WebSVN – Kolibri OS – Path Comparison – / – /programs/demos/ Rev 9236 and /programs/demos/ Rev 9237

Ignore whitespace Rev 9236 → Rev 9237

 /programs/demos/view3ds/3dmath.inc
 ,13 → 4,108
 vec_x equ 0
 vec_y equ 4
 vec_z equ 8
-; 3d point - triple integer word coordinate
-; vector   - triple float dword coordinate
-;----------------------in: --------------------------------
-;------------------------ esi - pointer to 1st 3d point ---
-;------------------------ edi - pointer to 2nd 3d point ---
-;------------------------ ebx - pointer to result vector --
-;---------------------- out : none ------------------------
+if 0  ; Ext >= SSE3
+calc_bounding_box:
+; in:
+;      xmm0  -  normal vector of ray
+;      xmm1  -  light origin
+; out:
+;      eax - axis aligned bounding boxes bit mask
+       .rmx     equ [ebp-36]
+       .nray    equ [ebp-64]
+       .origin  equ [ebp-80]
+       .dirfrac equ [ebp-96]
+       .nrayr   equ [ebp-112]
+       .originr equ [ebp-128]
+       .tmin    equ [ebp-132]
+       .tmax    equ [ebp-136]
+       push    ebp
+       mov     ebp,esp
+       and     ebp,-16
+       sub     esp,160
+       movss     xmm5,[rsscale]
+       shufps    xmm5,xmm1,0
+       movd      xmm2,[vect_x]
+       punpcklwd xmm2,[the_zero]
+       cvtdq2ps  xmm2,xmm2
+       subps     xmm1,xmm2
+       movaps    .origin,xmm1
+       mulps     xmm0,xmm5
+       movaps    .nray,xmm0
+       mov     esi,matrix
+       lea     edi,.rmx
+       call    reverse_mx_3x3
+;  in:  esi - ptr to points(normals], each point(normal) coeficient as dword
+;       edi - ptr to rotated points(normals)
+;       ebx - ptr to 3x3 (9 dwords, 36 bytes) rotation matrix
+;       ecx - number of points(normals)
+    ; reverse transform
+       lea    esi,.nray
+       lea    edi,.nrayr
+       lea    ebx,.rmx
+       mov    ecx,1
+       call   rotary
+       lea    esi,.origin
+       lea    edi,.originr
+       lea    ebx,.rmx
+       mov    ecx,1
+       call   rotary
+       xor      ecx,ecx
+       mov      ebx,aabb1
+       xor      eax,eax
+       rcpps    xmm7,.nrayr
+       movaps   .dirfrac,xmm7
+    .nx_aabb:
+       movaps   xmm5,[ebx]
+       movaps   xmm6,[ebx]
+       minps    xmm5,[the_zero]
+       maxps    xmm6,[the_zero]
+ ;  xmm5 - lb corner of AABB with minimal coordinates
+ ;  xmm6 - rt cor.   of AABB wit maximum coords
+       subps    xmm5,.originr
+       subps    xmm6,.originr
+       mulps    xmm5,.dirfrac   ; xmm5 - tx1, ty1
+       mulps    xmm6,.dirfrac   ; xmm6 - tx2, ty2
+       movaps   xmm1,xmm6
+       movaps   xmm2,xmm6
+       minps    xmm1,xmm5
+       maxps    xmm2,xmm5
+       movaps   xmm5,xmm1
+       movaps   xmm6,xmm2
+       shufps   xmm5,xmm5,11100001b
+       shufps   xmm6,xmm6,11100001b
+       maxss    xmm1,xmm5  ;t min
+       minss    xmm2,xmm6  ;t max
+       comiss   xmm2,xmm1
+       jb       .no_inter
+    .yes:
+       bts      eax,ecx
+     .no_inter:
+       add      ebx,16
+       inc      ecx
+       cmp      ecx,8
+       jne      .nx_aabb
+; out: eax - bit mask
+       add      esp,160
+       pop      ebp
+ret
+end if
 reverse_mx_3x3:
 ; esi - source matrix
 ; edi - desired reversed matrix
 ,6 → 236,13
   mov  esp,ebp
   pop  ebp
 ret
+; 3d point - triple integer word coordinate
+; vector   - triple float dword coordinate
+;----------------------in: --------------------------------
+;------------------------ esi - pointer to 1st 3d point ---
+;------------------------ edi - pointer to 2nd 3d point ---
+;------------------------ ebx - pointer to result vector --
+;---------------------- out : none ------------------------
 make_vector_r:
    if Ext < SSE2
 ,17 → 296,37
         fsubp   ;st1 ,st
         fstp    dword [ebx+vec_z]
 ret
+cross_aligned:
+      movaps  xmm0,[esi]
+      movaps  xmm1,[esi]
+      movaps  xmm2,[edi]
+      movaps  xmm3,[edi]
+      shufps  xmm0,xmm0,00001001b
+      shufps  xmm1,xmm1,00010010b
+      shufps  xmm2,xmm2,00010010b
+      shufps  xmm3,xmm3,00001001b
+      mulps   xmm0,xmm2
+      mulps   xmm1,xmm3
+      subps   xmm0,xmm1
+      movaps  [ebx],xmm0
+ret
 ;----------------------- in: ------------------------------
 ;---------------------------- edi - pointer to vector -----
 ;----------------------- out : none
 normalize_vector:
-if Ext >= SSE3
+if Ext >= SSE2
         movups  xmm0,[edi]
         andps   xmm0,[zero_hgst_dd]
         movups  xmm1,xmm0
         mulps   xmm0,xmm0
-        haddps  xmm0,xmm0
-        haddps  xmm0,xmm0
+        movhlps xmm2,xmm0
+        addps   xmm0,xmm2
+        movaps  xmm2,xmm0
+        shufps  xmm2,xmm2,11100101b
+        addps   xmm0,xmm2
+        shufps  xmm0,xmm0,0
+;        haddps  xmm0,xmm0
+;        haddps  xmm0,xmm0
         rsqrtps xmm0,xmm0
         mulps   xmm0,xmm1
         movlps  [edi],xmm0
 ,7 → 681,7
    ; packsdw xmm0,xmm0
    ; movq     [edi]
     fld    dword[esi]
-    fiadd  [vect_x]
+    fiadd  word[vect_x]
     fistp  word[edi]
     fld    dword[esi+4]
     fiadd  [vect_y]

/programs/demos/view3ds/3glass.inc
1,5 → 1,5
; Glass like rendering triangle by Maciej Guba.
; http://macgub.hekko.pl, macgub3@wp.pl
; http://macgub.co.pl, macgub3@wp.pl

ROUND2 equ 10
glass_tri:

 /programs/demos/view3ds/3glass_tex.inc
 ,762 → 1,762
-; Bilinear filtering, real Phongs shading and glass like parallel.
-; Thanks to authors of 3dica tutorial.
-; Implemented in FASM by Maciej Guba.
-; http://macgub.j.pl
-ROUND2 equ 10
-glass_tex_tri:
-;----Procedure render Phongs shaded triangle with z coord
-;----interpolation ( Catmull alghoritm ), each pixel is -
-;----covered by texture using bilinear filtering.--------
-;----I normalize normal vector in every pixel -----------
-;------------------in - eax - x1 shl 16 + y1 ------------
-;---------------------- ebx - x2 shl 16 + y2 ------------
-;---------------------- ecx - x3 shl 16 + y3 ------------
-;---------------------- esi - pointer to stencil buffer--
-;----------------------   filled with dd float variables-
-;---------------------- edi - pointer to screen buffer---
-;---------------------- edx - pointer to texture---------
-;---------------------- xmm0 - 1st normal vector --------
-;---------------------- xmm1 - 2cond normal vector ------
-;---------------------- xmm2 - 3rd normal vector --------
-;---------------------- xmm3 - normalized light vector --
-;---------------------- xmm4 - lo -> hi z1, z2, z3 coords
-;----------------------  as dwords floats ---------------
-;---------------------- xmm5 - lo -> hi y_min, y_max, ---
-;----------------------  x_min, x_max as dword integers -
-;---------------------- xmm6 - lo -> hi tx1, ty1, tx2, --
-;----------------------  ty2, tx3, ty3 as word, xres as--
-;----------------------  dword integers------------------
-;---------------------- stack - no parameters -----------
-;--------------------------------------------------------
-;----------------- procedure don't save registers !! ----
-  push  ebp
-  mov   ebp,esp
-  sub   esp,512
-  sub   ebp,16
-  and   ebp,0xfffffff0
-  .1_nv equ [ebp-16]
-  .2_nv equ [ebp-32]
-  .3_nv equ [ebp-48]
-  .l_v  equ [ebp-64]
-  .z3   equ [ebp-72]
-  .z2   equ [ebp-76]
-  .z1   equ [ebp-80]
-  .x1   equ [ebp-82]
-  .y1   equ [ebp-84]
-  .x2   equ [ebp-86]
-  .y2   equ [ebp-88]
-  .x3   equ [ebp-90]
-  .y3   equ [ebp-92]
-  .Zbuf equ [ebp-96]
-  .x_max equ  [ebp-100]
-  .x_min equ  [ebp-104]
-  .y_max equ  [ebp-108]
-  .y_min equ  [ebp-112]
-  .screen equ [ebp-116]
-  .dx12   equ [ebp-120]
-  .dx13   equ [ebp-124]
-  .dx23   equ [ebp-128]
-  .dn12   equ [ebp-144]
-  .dn13   equ [ebp-160]
-  .dn23   equ [ebp-176]
-  .cnv1   equ [ebp-192]  ; cur normal vectors
-  .cnv2   equ [ebp-208]
-  .x_res  equ [ebp-212]
-  .ty3    equ [ebp-214]
-  .tx3    equ [ebp-216]
-  .ty2    equ [ebp-218]
-  .tx2    equ [ebp-220]
-  .ty1    equ [ebp-222]
-  .tx1    equ [ebp-224]
-  .dz12   equ [ebp-232]
-  .dty12  equ [ebp-236]
-  .dtx12  equ [ebp-240]
-  .dz13   equ [ebp-248]
-  .dty13  equ [ebp-252]
-  .dtx13  equ [ebp-256]
-  .dz23   equ [ebp-264]
-  .dty23  equ [ebp-268]
-  .dtx23  equ [ebp-272]
-  .cz1    equ [ebp-280]
-  .cty1   equ [ebp-284]
-  .ctx1   equ [ebp-288]
-  .cz2    equ [ebp-296]
-  .cty2   equ [ebp-300]
-  .ctx2   equ [ebp-304]
-  .tx_ptr equ [ebp-308]
-      emms
-    ;  movd    .x_res,xmm7
- .sort3:                  ; sort triangle coordinates...
-       cmp     ax,bx
-       jle     .sort1
-       xchg    eax,ebx
-       shufps  xmm4,xmm4,11100001b
-       shufps  xmm6,xmm6,11100001b
-       movaps  xmm7,xmm0
-       movaps  xmm0,xmm1
-       movaps  xmm1,xmm7
- .sort1:
-       cmp      bx,cx
-       jle      .sort2
-       xchg     ebx,ecx
-       shufps   xmm4,xmm4,11011000b
-       shufps   xmm6,xmm6,11011000b
-       movaps   xmm7,xmm1
-       movaps   xmm1,xmm2
-       movaps   xmm2,xmm7
-       jmp .sort3
- .sort2:
- ;  movq    .tx1,xmm6
- ;  pshufd  xmm6,xmm6,01001110b
- ;  movd    .tx3,xmm6
-   movaps .tx1,xmm6
-   movaps .z1,xmm4
-   mov    .y1,eax
-   mov    .y2,ebx
-   mov    .y3,ecx
-   movdqa   .y_min,xmm5
-if 1                            ; check if at last only fragment
-   packssdw xmm5,xmm5       ; of triangle is in visable area
-   pshuflw  xmm5,xmm5,11011000b
-   movdqu   xmm7,.y3
-   movdqa   xmm6,xmm5
-   pshufd   xmm5,xmm5,0  ; xmm5 lo-hi -> broadcasted y_min, x_min
-   pshufd   xmm6,xmm6,01010101b ;xmm6 -> brd y_max x_max
-   movdqa   xmm4,xmm7
-   pcmpgtw  xmm7,xmm5
-   pcmpgtw  xmm4,xmm6
-   pxor     xmm7,xmm4
-   pmovmskb eax,xmm7
-   and      eax,0x00aaaaaa
-   or       eax,eax
-   jz       .rpt_loop2_end
-end if
-   movaps   .1_nv,xmm0
-   movaps   .2_nv,xmm1
-   movaps   .3_nv,xmm2
-   movaps   .l_v,xmm3
-   mov      .Zbuf,esi
-   mov      .screen,edi
-   mov      .tx_ptr,edx
-       mov      bx,.y2       ; calc deltas
-       sub      bx,.y1
-       jnz      .rpt_dx12_make
-       xorps    xmm7,xmm7
-       mov      dword .dx12,0
-       movaps   .dtx12,xmm7
-       movaps   .dn12,xmm7
-       jmp      .rpt_dx12_done
-  .rpt_dx12_make:
-       mov      ax,.x2
-       sub      ax,.x1
-       cwde
-       movsx    ebx,bx
-       shl      eax,ROUND2
-       cdq
-       idiv     ebx
-       mov      .dx12,eax
-       cvtsi2ss xmm6,ebx
-       shufps   xmm6,xmm6,0
-       movss    xmm5,.z2
-       subss    xmm5,.z1
-       divss    xmm5,xmm6
-       movss    .dz12,xmm5
-       movd     xmm0,.tx1
-       movd     xmm2,.tx2
-       pxor     xmm1,xmm1
-       punpcklwd xmm0,xmm1
-       punpcklwd xmm2,xmm1
-       psubd     xmm2,xmm0
-     ;  cvtdq2ps xmm0,xmm0
-       cvtdq2ps xmm2,xmm2
-;       movlps   .ctx1,xmm0
-;       movlps   .ctx2,xmm2
-     ;  subps    xmm2,xmm0
-       divps    xmm2,xmm6
-       movlps   .dtx12,xmm2
-       movaps   xmm0,.2_nv
-       subps    xmm0,.1_nv
-       divps    xmm0,xmm6
-       movaps   .dn12,xmm0
-   .rpt_dx12_done:
-       mov      bx,.y3       ; calc deltas
-       sub      bx,.y1
-       jnz      .rpt_dx13_make
-       xorps    xmm7,xmm7
-       mov      dword .dx13,0
-       movaps   .dtx13,xmm7
-       movaps   .dn13,xmm7
-       jmp      .rpt_dx13_done
-  .rpt_dx13_make:
-       mov      ax,.x3
-       sub      ax,.x1
-       cwde
-       movsx    ebx,bx
-       shl      eax,ROUND2
-       cdq
-       idiv     ebx
-       mov      .dx13,eax
-       cvtsi2ss xmm6,ebx
-       shufps   xmm6,xmm6,0
-       movss    xmm5,.z3
-       subss    xmm5,.z1
-       divss    xmm5,xmm6
-       movss    .dz13,xmm5
-       movd     xmm0,.tx1
-       movd     xmm2,.tx3
-       pxor     xmm1,xmm1
-       punpcklwd xmm0,xmm1
-       punpcklwd xmm2,xmm1
-       psubd    xmm2,xmm0
-     ;  cvtdq2ps xmm0,xmm0
-       cvtdq2ps xmm2,xmm2
-     ;  subps    xmm2,xmm0
-       divps    xmm2,xmm6
-       movlps   .dtx13,xmm2
-       movaps   xmm0,.3_nv
-       subps    xmm0,.1_nv
-       divps    xmm0,xmm6
-       movaps   .dn13,xmm0
-   .rpt_dx13_done:
-       mov      bx,.y3       ; calc deltas
-       sub      bx,.y2
-       jnz      .rpt_dx23_make
-       xorps    xmm7,xmm7
-       mov      dword .dx23,0
-       movaps   .dtx23,xmm7
-       movaps   .dn23,xmm7
-       jmp      .rpt_dx23_done
-  .rpt_dx23_make:
-       mov      ax,.x3
-       sub      ax,.x2
-       cwde
-       movsx    ebx,bx
-       shl      eax,ROUND2
-       cdq
-       idiv     ebx
-       mov      .dx23,eax
-       cvtsi2ss xmm6,ebx
-       shufps   xmm6,xmm6,0
-       movss    xmm5,.z3
-       subss    xmm5,.z2
-       divss    xmm5,xmm6
-       movss    .dz23,xmm5
-       movd     xmm0,.tx2
-       movd     xmm2,.tx3
-       pxor     xmm1,xmm1
-       punpcklwd xmm0,xmm1
-       punpcklwd xmm2,xmm1
-       psubd     xmm2,xmm0
-      ; cvtdq2ps xmm0,xmm0
-       cvtdq2ps xmm2,xmm2
-;       movlps   .ctx1,xmm0
-;       movlps   .ctx2,xmm2
-    ;   subps    xmm2,xmm0
-       divps    xmm2,xmm6
-       movlps   .dtx23,xmm2
-       movaps   xmm0,.3_nv
-       subps    xmm0,.2_nv
-       divps    xmm0,xmm6
-       movaps   .dn23,xmm0
-   .rpt_dx23_done:
-       movsx   eax,word .x1
-       shl     eax,ROUND2
-       mov     ebx,eax
-       mov     edx,.z1
-       movd    xmm1,.tx1
-       pxor    xmm2,xmm2
-       punpcklwd xmm1,xmm2
-       cvtdq2ps  xmm1,xmm1
-       mov     .cz1,edx
-       mov     .cz2,edx
-       movaps  xmm0,.1_nv
-       movlps  .ctx1,xmm1
-       movlps  .ctx2,xmm1
-       movaps  .cnv1,xmm0
-       movaps  .cnv2,xmm0
-     ;  mov     edx,.dx13
-     ;  cmp     edx,.dx12
-     ;  jg      .second_cause
-       movsx    ecx,word .y1
-       cmp      cx,.y2
-       jge      .rpt_loop1_end
-    .rpt_loop1:
-       pushad
-       movaps   xmm2,.y_min
-       movaps   xmm0,.cnv1
-       movaps   xmm1,.cnv2
-    ;   movlps   xmm3,.cz1 ; cz1, cz2 both
-       movaps   xmm3,.ctx1
-       movaps   xmm5,.ctx2
-       movaps   xmm4,.l_v
-       movd     xmm6,.x_res
-       sar      ebx,ROUND2
-       sar      eax,ROUND2
-       mov      edx,.tx_ptr
-       mov      edi,.screen
-       mov      esi,.Zbuf
-       call     glass_tex_line
-       popad
-       movaps   xmm0,.cnv1
-       movaps   xmm1,.cnv2
-     ;  movss    xmm2,.cz1
-     ;  movss    xmm3,.cz2
-       movaps   xmm2,.ctx1
-       movaps   xmm3,.ctx2
-       addps    xmm0,.dn13
-       addps    xmm1,.dn12
-       addps    xmm2,.dtx13
-       addps    xmm3,.dtx12
-       add      eax,.dx13
-       add      ebx,.dx12
-       movaps   .cnv1,xmm0
-       movaps   .cnv2,xmm1
-     ;  movss    .cz1,xmm2
-     ;  movss    .cz2,xmm3
-       movaps   .ctx1,xmm2
-       movaps   .ctx2,xmm3
-       add      ecx,1
-       cmp      cx,.y2
-       jl       .rpt_loop1
-   ;    jmp  .rpt_loop2_end
-   .rpt_loop1_end:
-       movsx    ecx,word .y2
-       cmp      cx,.y3
-       jge      .rpt_loop2_end
-       movsx    ebx,word .x2                    ; eax - cur x1
-       shl      ebx,ROUND2                 ; ebx - cur x2
-       push     dword .z2
-       pop      dword .cz2
-       movd     xmm1,.tx2
-       pxor     xmm2,xmm2
-       punpcklwd xmm1,xmm2
-       cvtdq2ps xmm1,xmm1
-       movlps   .ctx2,xmm1
-       movaps   xmm0,.2_nv
-       movaps   .cnv2,xmm0
-     .rpt_loop2:
-       pushad
-       movaps   xmm2,.y_min
-       movaps   xmm0,.cnv1
-       movaps   xmm1,.cnv2
-       movaps   xmm3,.ctx1
-       movaps   xmm5,.ctx2
-       movaps   xmm4,.l_v
-       sar      ebx,ROUND2
-       sar      eax,ROUND2
-       mov      edx,.tx_ptr
-       mov      edi,.screen
-       mov      esi,.Zbuf
-       movd     xmm6,.x_res
-       call     glass_tex_line
-       popad
-       movaps   xmm0,.cnv1
-       movaps   xmm1,.cnv2
-   ;    movss    xmm2,.cz1
-   ;    movss    xmm3,.cz2
-       movaps   xmm2,.ctx1
-       movaps   xmm3,.ctx2
-       addps    xmm0,.dn13
-       addps    xmm1,.dn23
-    ;   addss    xmm2,.dz13
-    ;   addss    xmm3,.dz23
-       addps    xmm2,.dtx13
-       addps    xmm3,.dtx23
-       add      eax,.dx13
-       add      ebx,.dx23
-       movaps   .cnv1,xmm0
-       movaps   .cnv2,xmm1
-       movaps   .ctx1,xmm2
-       movaps   .ctx2,xmm3
-     ;  movss    .cz1,xmm2
-     ;  movss    .cz2,xmm3
-       add      ecx,1
-       cmp      cx,.y3
-       jl       .rpt_loop2
-    .second_cause:  ;dx13 > dx12
-    .rpt_loop2_end:
-      add   esp,512
-      pop   ebp
-ret
-align 16
-glass_tex_line:
-; in:
-;    xmm0 - normal vector 1
-;    xmm1 - normal vect 2
-;    xmm3 - lo -> hi tx1, ty1, z1 coords as dwords float
-;    xmm5 - lo -> hi tx2, ty2, z2 coords as dwords float
-;    xmm2 - lo -> hi y_min, y_max, x_min, x_max
-;           as dword integers
-;    xmm4 - normalized light vector
-;    eax - x1
-;    ebx - x2
-;    ecx - y
-;    edi - screen buffer
-;    esi - stencil buffer filled with dd floats
-;    edx - texture pointer (handle)
-;    xmm6 - lowest dword x_res as integer
-   push  ebp
-   mov   ebp,esp
-   sub   esp,350
-   sub   ebp,16
-   and   ebp,0xfffffff0
- .n1 equ [ebp-16]
- .n2 equ [ebp-32]
- .lv equ [ebp-48]
- .lx1 equ [ebp-52]
- .lx2 equ [ebp-56]
-; .z2 equ [ebp-60]
-; .z1 equ [ebp-64]
- .screen equ [ebp-68]
- .zbuff  equ [ebp-72]
- .x_max  equ [ebp-74]
- .x_min  equ [ebp-76]
- .y_max  equ [ebp-78]
- .y_min  equ [ebp-80]
- .dn     equ [ebp-96]
- .x_res  equ [ebp-100]
- .y      equ [ebp-104]
- .cnv    equ [ebp-128]
- .z1     equ [ebp-136]
- .ty1    equ [ebp-140]
- .tx1    equ [ebp-144]
- .z2     equ [ebp-152]
- .ty2    equ [ebp-156]
- .tx2    equ [ebp-160]
- .cz     equ [ebp-168]
- .cty    equ [ebp-172]
- .ctx    equ [ebp-176]
- .dz     equ [ebp-184]
- .dty    equ [ebp-188]
- .dtx    equ [ebp-192]
- .yd     equ [ebp-196]
- .xd     equ [ebp-200]
- .yf     equ [ebp-204]
- .xf     equ [ebp-208]
- .w4     equ [ebp-212]
- .w3     equ [ebp-216]
- .w2     equ [ebp-220]
- .w1     equ [ebp-224]
- .p4     equ [ebp-228]
- .p3     equ [ebp-232]
- .p2     equ [ebp-236]
- .p1     equ [ebp-240]
- .tx_ptr equ [ebp-244]
-  ;      movaps  xmm7,xmm3
-  ;      movaps  xmm3,xmm5
-  ;      movaps  xmm5,xmm7
-        mov    .y,ecx
-        packssdw xmm2,xmm2
-    ;    movaps xmm7,xmm2
-    ;    movhps xmm2,[the_zero]
-    ;    pshuflw xmm2,xmm2,11111000b
-    ;    pshufd  xmm2,xmm2,11111100b
-   ;     movlps  xmm7,[the_zero]
-    ;    pshufhw xmm7,xmm7,11111111b
-     ;   movlps  xmm7,[the_zero]
-     ;   psrldq  xmm7,4
-     ;   por     xmm2,xmm7
-        movq   .y_min,xmm2
-        cmp    cx,.y_min
-        jl     .end_line
-        cmp    cx,.y_max
-        jge    .end_line          ;
-        cmp     eax,ebx
-        je      .end_line
-        jl      @f
-        xchg    eax,ebx
-        movaps  xmm7,xmm0
-        movaps  xmm0,xmm1
-        movaps  xmm1,xmm7
-        movaps  xmm7,xmm3
-        movaps  xmm3,xmm5
-        movaps  xmm5,xmm7
-   @@:
-        cmp     ax,.x_max
-        jge     .end_line
-        cmp     bx,.x_min
-        jle     .end_line
-        movaps  .lv,xmm4
-        movaps  .n1,xmm0
-        movaps  .n2,xmm1
-        mov     .lx1,eax
-        mov     .lx2,ebx
-        movaps  .tx1,xmm3
-        movaps  .tx2,xmm5
-        movd    .x_res,xmm6
-        mov     .tx_ptr,edx
-        sub     ebx,eax
-        cvtsi2ss xmm7,ebx
-        shufps  xmm7,xmm7,0
-        subps   xmm1,xmm0
-        divps   xmm1,xmm7
-        movaps  .dn,xmm1
-        subps   xmm5,xmm3
-        divps   xmm5,xmm7
-        movaps  .dtx,xmm5
-        mov      ebx,.lx1
-        cmp      bx,.x_min     ; clipping on function4
-        jge      @f
-        movzx    eax,word .x_min
-        sub      eax,ebx
-        cvtsi2ss xmm7,eax
-        shufps   xmm7,xmm7,0
-        mulps    xmm5,xmm7
-        mulps    xmm1,xmm7
-        addps    xmm5,.tx1
-        addps    xmm1,.n1
-        movsx    eax,word .x_min
-        movaps   .tx1,xmm5
-        movaps   .n1,xmm1
-        mov      dword .lx1,eax
-      @@:
-        movzx   eax,word .x_max
-        cmp     .lx2,eax
-        jl      @f
-        mov     .lx2,eax
-      @@:
-        mov     eax,.x_res
-        mul     dword .y
-        add     eax,.lx1
-        shl     eax,2
-        add     edi,eax
-        add     esi,eax
-        mov     ecx,.lx2
-        sub     ecx,.lx1
-   ;     movaps  xmm0,.n1
-        movaps  xmm2,.tx1
-      ;  xorps   xmm1,xmm1
-align 16
-   .ddraw:
-     ;   movhlps  xmm7,xmm2
-     ;   cmpnltss xmm7,dword[esi]
-     ;   movd     eax,xmm7
-     ;   or       eax,eax
-     ;   jnz      .skip
-        xorps    xmm5,xmm5
-   ;     movhlps  xmm7,xmm2
-   ;     movss    [esi],xmm7
-        movaps   xmm7,.n1  ;xmm0
-        mulps    xmm7,xmm7 ; normalize
-        haddps   xmm7,xmm7
-        haddps   xmm7,xmm7
-        rsqrtps  xmm7,xmm7
-        mulps    xmm7,.n1   ;xmm0
- ;       andps    xmm7,[abs_z_coof]
-        movaps   .cnv,xmm7
-        movaps   xmm6,xmm2
-        minps    xmm6,[tex_m2]  ;    float  TEX_X-2,TEX_Y-2
-        cvttps2dq xmm7,xmm6
-        cvtdq2ps xmm4,xmm7
-        subps    xmm6,xmm4
-        movlps   .xf,xmm6
-     ;    movaps   xmm5,.lv
-        mov      eax,lights_aligned   ; global
-     align 16
-      .again_col:
-        movaps   xmm0,[eax] ; calc multple lights
-        mulps    xmm0,.cnv  ;.lv  ; last dword should be zeroed
-        haddps   xmm0,xmm0
-        haddps   xmm0,xmm0
-   ;     andps    xmm0,[abs_val]  ;calc absolute value
-if 1
-        ; stencil
-        movhlps   xmm6,xmm2
-        movhlps   xmm4,xmm2
-        addss     xmm6,[aprox]
-        subss     xmm4,[aprox]
-        cmpnltss  xmm6,dword[esi]
-        cmpnltss  xmm4,dword[esi]
-        xorps     xmm6,xmm4
-        xorps     xmm4,xmm4
-        movd      ebx,xmm6
-        cmp       ebx,-1
-        jne       .no_reflective
-end if
-        movaps    xmm4,xmm0
-        mulps     xmm4,xmm4
-        mulps     xmm4,xmm4
-        mulps     xmm4,xmm4
-        mulps     xmm4,xmm4
-        mulps     xmm4,[eax+48]
-     .no_reflective:
-        maxps    xmm0,[the_zero]
- ;       movaps   xmm1,xmm0
-        mulps    xmm0,[eax+16]
-        addps    xmm4,xmm0
-        addps    xmm4,[eax+32]
-        maxps    xmm5,xmm4
-        add      eax,64
-        cmp      eax,lights_aligned_end
-        jnz      .again_col
-        minps    xmm5,[mask_255f]
-          ; texture coords work
-        movd     eax,xmm7
-        psrldq   xmm7,4
-        movd     ebx,xmm7
-        shl      ebx,TEX_SHIFT
-        add      eax,ebx
-        lea      eax,[eax*3]
-        add      eax,.tx_ptr
-        mov      ebx,eax
-        add      ebx,TEX_X*3
-        movd     xmm7,[eax]
-        movd     xmm6,[eax+3]
-        movd     xmm4,[ebx]
-        movd     xmm3,[ebx+3]
-        punpcklbw xmm7,xmm6 ;xmm7 r1 r2 g1 g2 b1 b2
-        punpcklbw xmm4,xmm3 ;xmm4 r3 r4 g3 g4 b3 b4
-        punpcklwd xmm7,xmm4 ;xmm7 r1 r2 r3 r4 g1 g2 g3 g4 ...
-        movdqa    xmm6,xmm7
-        movdqa    xmm4,xmm7
-        psrldq    xmm6,4
-        psrldq    xmm4,8
-        punpcklbw xmm7,[the_zero]  ; broadcasted 0
-        punpcklbw xmm6,[the_zero]
-        punpcklbw xmm4,[the_zero]
-        punpcklwd xmm7,[the_zero]
-        punpcklwd xmm6,[the_zero]
-        punpcklwd xmm4,[the_zero]
-     ; calc w .........
-        movlps   xmm3,[the_one]  ;  broadcasted dword 1.0
-        cvtdq2ps  xmm7,xmm7
-        subps    xmm3,.xf
-        cvtdq2ps  xmm6,xmm6
-        movhps   xmm3,.xf
-        cvtdq2ps  xmm4,xmm4
-        movaps   xmm1,xmm3  ; 1-xf, 1-yf, xf, yf
-        shufps   xmm3,xmm3,10001000b
-        shufps   xmm1,xmm1,11110101b
-        mulps    xmm3,xmm1
-        mulps    xmm7,xmm3
-        mulps    xmm6,xmm3
-        mulps    xmm4,xmm3
-        haddps   xmm7,xmm7  ; r
-        haddps   xmm6,xmm6  ; g
-        haddps   xmm4,xmm4  ; b
-        haddps   xmm7,xmm7  ; r
-        haddps   xmm6,xmm6  ; g
-        haddps   xmm4,xmm4  ; b
-        movlhps  xmm7,xmm6
-        shufps   xmm7,xmm7,11101000b
-        movlhps  xmm7,xmm4
-        mulps    xmm5,xmm7
-        cvtps2dq xmm5,xmm5
-        psrld    xmm5,8
-        movd     xmm6,[edi]
-        packssdw xmm5,xmm5
-        packuswb xmm5,xmm5
-        paddusb  xmm5,xmm6
-        movd     [edi],xmm5
-     .skip:
-        add      edi,4
-        add      esi,4
-      ;  addps    xmm0,.dn
-        movaps   xmm0,.n1     ; cur normal
-        addps    xmm0,.dn
-        addps    xmm2,.dtx
-        movaps   .n1,xmm0
-        sub      ecx,1
-        jnz      .ddraw
-  .end_line:
-        add      esp,350
-        pop      ebp
-ret
+; Bilinear filtering, real Phongs shading and glass like parallel.
+; Thanks to authors of 3dica tutorial.
+; Implemented in FASM by Maciej Guba.
+; http://macgub.co.pl
+ROUND2 equ 10
+glass_tex_tri:
+;----Procedure render Phongs shaded triangle with z coord
+;----interpolation ( Catmull alghoritm ), each pixel is -
+;----covered by texture using bilinear filtering.--------
+;----I normalize normal vector in every pixel -----------
+;------------------in - eax - x1 shl 16 + y1 ------------
+;---------------------- ebx - x2 shl 16 + y2 ------------
+;---------------------- ecx - x3 shl 16 + y3 ------------
+;---------------------- esi - pointer to stencil buffer--
+;----------------------   filled with dd float variables-
+;---------------------- edi - pointer to screen buffer---
+;---------------------- edx - pointer to texture---------
+;---------------------- xmm0 - 1st normal vector --------
+;---------------------- xmm1 - 2cond normal vector ------
+;---------------------- xmm2 - 3rd normal vector --------
+;---------------------- xmm3 - normalized light vector --
+;---------------------- xmm4 - lo -> hi z1, z2, z3 coords
+;----------------------  as dwords floats ---------------
+;---------------------- xmm5 - lo -> hi y_min, y_max, ---
+;----------------------  x_min, x_max as dword integers -
+;---------------------- xmm6 - lo -> hi tx1, ty1, tx2, --
+;----------------------  ty2, tx3, ty3 as word, xres as--
+;----------------------  dword integers------------------
+;---------------------- stack - no parameters -----------
+;--------------------------------------------------------
+;----------------- procedure don't save registers !! ----
+  push  ebp
+  mov   ebp,esp
+  sub   esp,512
+  sub   ebp,16
+  and   ebp,0xfffffff0
+  .1_nv equ [ebp-16]
+  .2_nv equ [ebp-32]
+  .3_nv equ [ebp-48]
+  .l_v  equ [ebp-64]
+  .z3   equ [ebp-72]
+  .z2   equ [ebp-76]
+  .z1   equ [ebp-80]
+  .x1   equ [ebp-82]
+  .y1   equ [ebp-84]
+  .x2   equ [ebp-86]
+  .y2   equ [ebp-88]
+  .x3   equ [ebp-90]
+  .y3   equ [ebp-92]
+  .Zbuf equ [ebp-96]
+  .x_max equ  [ebp-100]
+  .x_min equ  [ebp-104]
+  .y_max equ  [ebp-108]
+  .y_min equ  [ebp-112]
+  .screen equ [ebp-116]
+  .dx12   equ [ebp-120]
+  .dx13   equ [ebp-124]
+  .dx23   equ [ebp-128]
+  .dn12   equ [ebp-144]
+  .dn13   equ [ebp-160]
+  .dn23   equ [ebp-176]
+  .cnv1   equ [ebp-192]  ; cur normal vectors
+  .cnv2   equ [ebp-208]
+  .x_res  equ [ebp-212]
+  .ty3    equ [ebp-214]
+  .tx3    equ [ebp-216]
+  .ty2    equ [ebp-218]
+  .tx2    equ [ebp-220]
+  .ty1    equ [ebp-222]
+  .tx1    equ [ebp-224]
+  .dz12   equ [ebp-232]
+  .dty12  equ [ebp-236]
+  .dtx12  equ [ebp-240]
+  .dz13   equ [ebp-248]
+  .dty13  equ [ebp-252]
+  .dtx13  equ [ebp-256]
+  .dz23   equ [ebp-264]
+  .dty23  equ [ebp-268]
+  .dtx23  equ [ebp-272]
+  .cz1    equ [ebp-280]
+  .cty1   equ [ebp-284]
+  .ctx1   equ [ebp-288]
+  .cz2    equ [ebp-296]
+  .cty2   equ [ebp-300]
+  .ctx2   equ [ebp-304]
+  .tx_ptr equ [ebp-308]
+      emms
+    ;  movd    .x_res,xmm7
+ .sort3:                  ; sort triangle coordinates...
+       cmp     ax,bx
+       jle     .sort1
+       xchg    eax,ebx
+       shufps  xmm4,xmm4,11100001b
+       shufps  xmm6,xmm6,11100001b
+       movaps  xmm7,xmm0
+       movaps  xmm0,xmm1
+       movaps  xmm1,xmm7
+ .sort1:
+       cmp      bx,cx
+       jle      .sort2
+       xchg     ebx,ecx
+       shufps   xmm4,xmm4,11011000b
+       shufps   xmm6,xmm6,11011000b
+       movaps   xmm7,xmm1
+       movaps   xmm1,xmm2
+       movaps   xmm2,xmm7
+       jmp .sort3
+ .sort2:
+ ;  movq    .tx1,xmm6
+ ;  pshufd  xmm6,xmm6,01001110b
+ ;  movd    .tx3,xmm6
+   movaps .tx1,xmm6
+   movaps .z1,xmm4
+   mov    .y1,eax
+   mov    .y2,ebx
+   mov    .y3,ecx
+   movdqa   .y_min,xmm5
+if 1                            ; check if at last only fragment
+   packssdw xmm5,xmm5       ; of triangle is in visable area
+   pshuflw  xmm5,xmm5,11011000b
+   movdqu   xmm7,.y3
+   movdqa   xmm6,xmm5
+   pshufd   xmm5,xmm5,0  ; xmm5 lo-hi -> broadcasted y_min, x_min
+   pshufd   xmm6,xmm6,01010101b ;xmm6 -> brd y_max x_max
+   movdqa   xmm4,xmm7
+   pcmpgtw  xmm7,xmm5
+   pcmpgtw  xmm4,xmm6
+   pxor     xmm7,xmm4
+   pmovmskb eax,xmm7
+   and      eax,0x00aaaaaa
+   or       eax,eax
+   jz       .rpt_loop2_end
+end if
+   movaps   .1_nv,xmm0
+   movaps   .2_nv,xmm1
+   movaps   .3_nv,xmm2
+   movaps   .l_v,xmm3
+   mov      .Zbuf,esi
+   mov      .screen,edi
+   mov      .tx_ptr,edx
+       mov      bx,.y2       ; calc deltas
+       sub      bx,.y1
+       jnz      .rpt_dx12_make
+       xorps    xmm7,xmm7
+       mov      dword .dx12,0
+       movaps   .dtx12,xmm7
+       movaps   .dn12,xmm7
+       jmp      .rpt_dx12_done
+  .rpt_dx12_make:
+       mov      ax,.x2
+       sub      ax,.x1
+       cwde
+       movsx    ebx,bx
+       shl      eax,ROUND2
+       cdq
+       idiv     ebx
+       mov      .dx12,eax
+       cvtsi2ss xmm6,ebx
+       shufps   xmm6,xmm6,0
+       movss    xmm5,.z2
+       subss    xmm5,.z1
+       divss    xmm5,xmm6
+       movss    .dz12,xmm5
+       movd     xmm0,.tx1
+       movd     xmm2,.tx2
+       pxor     xmm1,xmm1
+       punpcklwd xmm0,xmm1
+       punpcklwd xmm2,xmm1
+       psubd     xmm2,xmm0
+     ;  cvtdq2ps xmm0,xmm0
+       cvtdq2ps xmm2,xmm2
+;       movlps   .ctx1,xmm0
+;       movlps   .ctx2,xmm2
+     ;  subps    xmm2,xmm0
+       divps    xmm2,xmm6
+       movlps   .dtx12,xmm2
+       movaps   xmm0,.2_nv
+       subps    xmm0,.1_nv
+       divps    xmm0,xmm6
+       movaps   .dn12,xmm0
+   .rpt_dx12_done:
+       mov      bx,.y3       ; calc deltas
+       sub      bx,.y1
+       jnz      .rpt_dx13_make
+       xorps    xmm7,xmm7
+       mov      dword .dx13,0
+       movaps   .dtx13,xmm7
+       movaps   .dn13,xmm7
+       jmp      .rpt_dx13_done
+  .rpt_dx13_make:
+       mov      ax,.x3
+       sub      ax,.x1
+       cwde
+       movsx    ebx,bx
+       shl      eax,ROUND2
+       cdq
+       idiv     ebx
+       mov      .dx13,eax
+       cvtsi2ss xmm6,ebx
+       shufps   xmm6,xmm6,0
+       movss    xmm5,.z3
+       subss    xmm5,.z1
+       divss    xmm5,xmm6
+       movss    .dz13,xmm5
+       movd     xmm0,.tx1
+       movd     xmm2,.tx3
+       pxor     xmm1,xmm1
+       punpcklwd xmm0,xmm1
+       punpcklwd xmm2,xmm1
+       psubd    xmm2,xmm0
+     ;  cvtdq2ps xmm0,xmm0
+       cvtdq2ps xmm2,xmm2
+     ;  subps    xmm2,xmm0
+       divps    xmm2,xmm6
+       movlps   .dtx13,xmm2
+       movaps   xmm0,.3_nv
+       subps    xmm0,.1_nv
+       divps    xmm0,xmm6
+       movaps   .dn13,xmm0
+   .rpt_dx13_done:
+       mov      bx,.y3       ; calc deltas
+       sub      bx,.y2
+       jnz      .rpt_dx23_make
+       xorps    xmm7,xmm7
+       mov      dword .dx23,0
+       movaps   .dtx23,xmm7
+       movaps   .dn23,xmm7
+       jmp      .rpt_dx23_done
+  .rpt_dx23_make:
+       mov      ax,.x3
+       sub      ax,.x2
+       cwde
+       movsx    ebx,bx
+       shl      eax,ROUND2
+       cdq
+       idiv     ebx
+       mov      .dx23,eax
+       cvtsi2ss xmm6,ebx
+       shufps   xmm6,xmm6,0
+       movss    xmm5,.z3
+       subss    xmm5,.z2
+       divss    xmm5,xmm6
+       movss    .dz23,xmm5
+       movd     xmm0,.tx2
+       movd     xmm2,.tx3
+       pxor     xmm1,xmm1
+       punpcklwd xmm0,xmm1
+       punpcklwd xmm2,xmm1
+       psubd     xmm2,xmm0
+      ; cvtdq2ps xmm0,xmm0
+       cvtdq2ps xmm2,xmm2
+;       movlps   .ctx1,xmm0
+;       movlps   .ctx2,xmm2
+    ;   subps    xmm2,xmm0
+       divps    xmm2,xmm6
+       movlps   .dtx23,xmm2
+       movaps   xmm0,.3_nv
+       subps    xmm0,.2_nv
+       divps    xmm0,xmm6
+       movaps   .dn23,xmm0
+   .rpt_dx23_done:
+       movsx   eax,word .x1
+       shl     eax,ROUND2
+       mov     ebx,eax
+       mov     edx,.z1
+       movd    xmm1,.tx1
+       pxor    xmm2,xmm2
+       punpcklwd xmm1,xmm2
+       cvtdq2ps  xmm1,xmm1
+       mov     .cz1,edx
+       mov     .cz2,edx
+       movaps  xmm0,.1_nv
+       movlps  .ctx1,xmm1
+       movlps  .ctx2,xmm1
+       movaps  .cnv1,xmm0
+       movaps  .cnv2,xmm0
+     ;  mov     edx,.dx13
+     ;  cmp     edx,.dx12
+     ;  jg      .second_cause
+       movsx    ecx,word .y1
+       cmp      cx,.y2
+       jge      .rpt_loop1_end
+    .rpt_loop1:
+       pushad
+       movaps   xmm2,.y_min
+       movaps   xmm0,.cnv1
+       movaps   xmm1,.cnv2
+    ;   movlps   xmm3,.cz1 ; cz1, cz2 both
+       movaps   xmm3,.ctx1
+       movaps   xmm5,.ctx2
+       movaps   xmm4,.l_v
+       movd     xmm6,.x_res
+       sar      ebx,ROUND2
+       sar      eax,ROUND2
+       mov      edx,.tx_ptr
+       mov      edi,.screen
+       mov      esi,.Zbuf
+       call     glass_tex_line
+       popad
+       movaps   xmm0,.cnv1
+       movaps   xmm1,.cnv2
+     ;  movss    xmm2,.cz1
+     ;  movss    xmm3,.cz2
+       movaps   xmm2,.ctx1
+       movaps   xmm3,.ctx2
+       addps    xmm0,.dn13
+       addps    xmm1,.dn12
+       addps    xmm2,.dtx13
+       addps    xmm3,.dtx12
+       add      eax,.dx13
+       add      ebx,.dx12
+       movaps   .cnv1,xmm0
+       movaps   .cnv2,xmm1
+     ;  movss    .cz1,xmm2
+     ;  movss    .cz2,xmm3
+       movaps   .ctx1,xmm2
+       movaps   .ctx2,xmm3
+       add      ecx,1
+       cmp      cx,.y2
+       jl       .rpt_loop1
+   ;    jmp  .rpt_loop2_end
+   .rpt_loop1_end:
+       movsx    ecx,word .y2
+       cmp      cx,.y3
+       jge      .rpt_loop2_end
+       movsx    ebx,word .x2                    ; eax - cur x1
+       shl      ebx,ROUND2                 ; ebx - cur x2
+       push     dword .z2
+       pop      dword .cz2
+       movd     xmm1,.tx2
+       pxor     xmm2,xmm2
+       punpcklwd xmm1,xmm2
+       cvtdq2ps xmm1,xmm1
+       movlps   .ctx2,xmm1
+       movaps   xmm0,.2_nv
+       movaps   .cnv2,xmm0
+     .rpt_loop2:
+       pushad
+       movaps   xmm2,.y_min
+       movaps   xmm0,.cnv1
+       movaps   xmm1,.cnv2
+       movaps   xmm3,.ctx1
+       movaps   xmm5,.ctx2
+       movaps   xmm4,.l_v
+       sar      ebx,ROUND2
+       sar      eax,ROUND2
+       mov      edx,.tx_ptr
+       mov      edi,.screen
+       mov      esi,.Zbuf
+       movd     xmm6,.x_res
+       call     glass_tex_line
+       popad
+       movaps   xmm0,.cnv1
+       movaps   xmm1,.cnv2
+   ;    movss    xmm2,.cz1
+   ;    movss    xmm3,.cz2
+       movaps   xmm2,.ctx1
+       movaps   xmm3,.ctx2
+       addps    xmm0,.dn13
+       addps    xmm1,.dn23
+    ;   addss    xmm2,.dz13
+    ;   addss    xmm3,.dz23
+       addps    xmm2,.dtx13
+       addps    xmm3,.dtx23
+       add      eax,.dx13
+       add      ebx,.dx23
+       movaps   .cnv1,xmm0
+       movaps   .cnv2,xmm1
+       movaps   .ctx1,xmm2
+       movaps   .ctx2,xmm3
+     ;  movss    .cz1,xmm2
+     ;  movss    .cz2,xmm3
+       add      ecx,1
+       cmp      cx,.y3
+       jl       .rpt_loop2
+    .second_cause:  ;dx13 > dx12
+    .rpt_loop2_end:
+      add   esp,512
+      pop   ebp
+ret
+align 16
+glass_tex_line:
+; in:
+;    xmm0 - normal vector 1
+;    xmm1 - normal vect 2
+;    xmm3 - lo -> hi tx1, ty1, z1 coords as dwords float
+;    xmm5 - lo -> hi tx2, ty2, z2 coords as dwords float
+;    xmm2 - lo -> hi y_min, y_max, x_min, x_max
+;           as dword integers
+;    xmm4 - normalized light vector
+;    eax - x1
+;    ebx - x2
+;    ecx - y
+;    edi - screen buffer
+;    esi - stencil buffer filled with dd floats
+;    edx - texture pointer (handle)
+;    xmm6 - lowest dword x_res as integer
+   push  ebp
+   mov   ebp,esp
+   sub   esp,350
+   sub   ebp,16
+   and   ebp,0xfffffff0
+ .n1 equ [ebp-16]
+ .n2 equ [ebp-32]
+ .lv equ [ebp-48]
+ .lx1 equ [ebp-52]
+ .lx2 equ [ebp-56]
+; .z2 equ [ebp-60]
+; .z1 equ [ebp-64]
+ .screen equ [ebp-68]
+ .zbuff  equ [ebp-72]
+ .x_max  equ [ebp-74]
+ .x_min  equ [ebp-76]
+ .y_max  equ [ebp-78]
+ .y_min  equ [ebp-80]
+ .dn     equ [ebp-96]
+ .x_res  equ [ebp-100]
+ .y      equ [ebp-104]
+ .cnv    equ [ebp-128]
+ .z1     equ [ebp-136]
+ .ty1    equ [ebp-140]
+ .tx1    equ [ebp-144]
+ .z2     equ [ebp-152]
+ .ty2    equ [ebp-156]
+ .tx2    equ [ebp-160]
+ .cz     equ [ebp-168]
+ .cty    equ [ebp-172]
+ .ctx    equ [ebp-176]
+ .dz     equ [ebp-184]
+ .dty    equ [ebp-188]
+ .dtx    equ [ebp-192]
+ .yd     equ [ebp-196]
+ .xd     equ [ebp-200]
+ .yf     equ [ebp-204]
+ .xf     equ [ebp-208]
+ .w4     equ [ebp-212]
+ .w3     equ [ebp-216]
+ .w2     equ [ebp-220]
+ .w1     equ [ebp-224]
+ .p4     equ [ebp-228]
+ .p3     equ [ebp-232]
+ .p2     equ [ebp-236]
+ .p1     equ [ebp-240]
+ .tx_ptr equ [ebp-244]
+  ;      movaps  xmm7,xmm3
+  ;      movaps  xmm3,xmm5
+  ;      movaps  xmm5,xmm7
+        mov    .y,ecx
+        packssdw xmm2,xmm2
+    ;    movaps xmm7,xmm2
+    ;    movhps xmm2,[the_zero]
+    ;    pshuflw xmm2,xmm2,11111000b
+    ;    pshufd  xmm2,xmm2,11111100b
+   ;     movlps  xmm7,[the_zero]
+    ;    pshufhw xmm7,xmm7,11111111b
+     ;   movlps  xmm7,[the_zero]
+     ;   psrldq  xmm7,4
+     ;   por     xmm2,xmm7
+        movq   .y_min,xmm2
+        cmp    cx,.y_min
+        jl     .end_line
+        cmp    cx,.y_max
+        jge    .end_line          ;
+        cmp     eax,ebx
+        je      .end_line
+        jl      @f
+        xchg    eax,ebx
+        movaps  xmm7,xmm0
+        movaps  xmm0,xmm1
+        movaps  xmm1,xmm7
+        movaps  xmm7,xmm3
+        movaps  xmm3,xmm5
+        movaps  xmm5,xmm7
+   @@:
+        cmp     ax,.x_max
+        jge     .end_line
+        cmp     bx,.x_min
+        jle     .end_line
+        movaps  .lv,xmm4
+        movaps  .n1,xmm0
+        movaps  .n2,xmm1
+        mov     .lx1,eax
+        mov     .lx2,ebx
+        movaps  .tx1,xmm3
+        movaps  .tx2,xmm5
+        movd    .x_res,xmm6
+        mov     .tx_ptr,edx
+        sub     ebx,eax
+        cvtsi2ss xmm7,ebx
+        shufps  xmm7,xmm7,0
+        subps   xmm1,xmm0
+        divps   xmm1,xmm7
+        movaps  .dn,xmm1
+        subps   xmm5,xmm3
+        divps   xmm5,xmm7
+        movaps  .dtx,xmm5
+        mov      ebx,.lx1
+        cmp      bx,.x_min     ; clipping on function4
+        jge      @f
+        movzx    eax,word .x_min
+        sub      eax,ebx
+        cvtsi2ss xmm7,eax
+        shufps   xmm7,xmm7,0
+        mulps    xmm5,xmm7
+        mulps    xmm1,xmm7
+        addps    xmm5,.tx1
+        addps    xmm1,.n1
+        movsx    eax,word .x_min
+        movaps   .tx1,xmm5
+        movaps   .n1,xmm1
+        mov      dword .lx1,eax
+      @@:
+        movzx   eax,word .x_max
+        cmp     .lx2,eax
+        jl      @f
+        mov     .lx2,eax
+      @@:
+        mov     eax,.x_res
+        mul     dword .y
+        add     eax,.lx1
+        shl     eax,2
+        add     edi,eax
+        add     esi,eax
+        mov     ecx,.lx2
+        sub     ecx,.lx1
+   ;     movaps  xmm0,.n1
+        movaps  xmm2,.tx1
+      ;  xorps   xmm1,xmm1
+align 16
+   .ddraw:
+     ;   movhlps  xmm7,xmm2
+     ;   cmpnltss xmm7,dword[esi]
+     ;   movd     eax,xmm7
+     ;   or       eax,eax
+     ;   jnz      .skip
+        xorps    xmm5,xmm5
+   ;     movhlps  xmm7,xmm2
+   ;     movss    [esi],xmm7
+        movaps   xmm7,.n1  ;xmm0
+        mulps    xmm7,xmm7 ; normalize
+        haddps   xmm7,xmm7
+        haddps   xmm7,xmm7
+        rsqrtps  xmm7,xmm7
+        mulps    xmm7,.n1   ;xmm0
+ ;       andps    xmm7,[abs_z_coof]
+        movaps   .cnv,xmm7
+        movaps   xmm6,xmm2
+        minps    xmm6,[tex_m2]  ;    float  TEX_X-2,TEX_Y-2
+        cvttps2dq xmm7,xmm6
+        cvtdq2ps xmm4,xmm7
+        subps    xmm6,xmm4
+        movlps   .xf,xmm6
+     ;    movaps   xmm5,.lv
+        mov      eax,lights_aligned   ; global
+     align 16
+      .again_col:
+        movaps   xmm0,[eax] ; calc multple lights
+        mulps    xmm0,.cnv  ;.lv  ; last dword should be zeroed
+        haddps   xmm0,xmm0
+        haddps   xmm0,xmm0
+   ;     andps    xmm0,[abs_val]  ;calc absolute value
+if 1
+        ; stencil
+        movhlps   xmm6,xmm2
+        movhlps   xmm4,xmm2
+        addss     xmm6,[aprox]
+        subss     xmm4,[aprox]
+        cmpnltss  xmm6,dword[esi]
+        cmpnltss  xmm4,dword[esi]
+        xorps     xmm6,xmm4
+        xorps     xmm4,xmm4
+        movd      ebx,xmm6
+        cmp       ebx,-1
+        jne       .no_reflective
+end if
+        movaps    xmm4,xmm0
+        mulps     xmm4,xmm4
+        mulps     xmm4,xmm4
+        mulps     xmm4,xmm4
+        mulps     xmm4,xmm4
+        mulps     xmm4,[eax+48]
+     .no_reflective:
+        maxps    xmm0,[the_zero]
+ ;       movaps   xmm1,xmm0
+        mulps    xmm0,[eax+16]
+        addps    xmm4,xmm0
+        addps    xmm4,[eax+32]
+        maxps    xmm5,xmm4
+        add      eax,64
+        cmp      eax,lights_aligned_end
+        jnz      .again_col
+        minps    xmm5,[mask_255f]
+          ; texture coords work
+        movd     eax,xmm7
+        psrldq   xmm7,4
+        movd     ebx,xmm7
+        shl      ebx,TEX_SHIFT
+        add      eax,ebx
+        lea      eax,[eax*3]
+        add      eax,.tx_ptr
+        mov      ebx,eax
+        add      ebx,TEX_X*3
+        movd     xmm7,[eax]
+        movd     xmm6,[eax+3]
+        movd     xmm4,[ebx]
+        movd     xmm3,[ebx+3]
+        punpcklbw xmm7,xmm6 ;xmm7 r1 r2 g1 g2 b1 b2
+        punpcklbw xmm4,xmm3 ;xmm4 r3 r4 g3 g4 b3 b4
+        punpcklwd xmm7,xmm4 ;xmm7 r1 r2 r3 r4 g1 g2 g3 g4 ...
+        movdqa    xmm6,xmm7
+        movdqa    xmm4,xmm7
+        psrldq    xmm6,4
+        psrldq    xmm4,8
+        punpcklbw xmm7,[the_zero]  ; broadcasted 0
+        punpcklbw xmm6,[the_zero]
+        punpcklbw xmm4,[the_zero]
+        punpcklwd xmm7,[the_zero]
+        punpcklwd xmm6,[the_zero]
+        punpcklwd xmm4,[the_zero]
+     ; calc w .........
+        movlps   xmm3,[the_one]  ;  broadcasted dword 1.0
+        cvtdq2ps  xmm7,xmm7
+        subps    xmm3,.xf
+        cvtdq2ps  xmm6,xmm6
+        movhps   xmm3,.xf
+        cvtdq2ps  xmm4,xmm4
+        movaps   xmm1,xmm3  ; 1-xf, 1-yf, xf, yf
+        shufps   xmm3,xmm3,10001000b
+        shufps   xmm1,xmm1,11110101b
+        mulps    xmm3,xmm1
+        mulps    xmm7,xmm3
+        mulps    xmm6,xmm3
+        mulps    xmm4,xmm3
+        haddps   xmm7,xmm7  ; r
+        haddps   xmm6,xmm6  ; g
+        haddps   xmm4,xmm4  ; b
+        haddps   xmm7,xmm7  ; r
+        haddps   xmm6,xmm6  ; g
+        haddps   xmm4,xmm4  ; b
+        movlhps  xmm7,xmm6
+        shufps   xmm7,xmm7,11101000b
+        movlhps  xmm7,xmm4
+        mulps    xmm5,xmm7
+        cvtps2dq xmm5,xmm5
+        psrld    xmm5,8
+        movd     xmm6,[edi]
+        packssdw xmm5,xmm5
+        packuswb xmm5,xmm5
+        paddusb  xmm5,xmm6
+        movd     [edi],xmm5
+     .skip:
+        add      edi,4
+        add      esi,4
+      ;  addps    xmm0,.dn
+        movaps   xmm0,.n1     ; cur normal
+        addps    xmm0,.dn
+        addps    xmm2,.dtx
+        movaps   .n1,xmm0
+        sub      ecx,1
+        jnz      .ddraw
+  .end_line:
+        add      esp,350
+        pop      ebp
+ret

 /programs/demos/view3ds/3r_phg.inc
 ,528 → 1,528
-; Real Phong's shading implemented if flat assembler
-; by Maciej Guba.
-; http://macgub.vxm.pl
-ROUND2 equ 10
-real_phong_tri_z:
-;----procedure render Phongs shaded triangle with z coord
-;----interpolation ( Catmull alghoritm )-----------------
-;----I normalize normal vector in every pixel -----------
-;------------------in - eax - x1 shl 16 + y1 ------------
-;---------------------- ebx - x2 shl 16 + y2 ------------
-;---------------------- ecx - x3 shl 16 + y3 ------------
-;---------------------- esi - pointer to Z-buffer filled-
-;----------------------   with dd float variables--------
-;---------------------- edi - pointer to screen buffer---
-;---------------------- xmm0 - 1st normal vector --------
-;---------------------- xmm1 - 2cond normal vector ------
-;---------------------- xmm2 - 3rd normal vector --------
-;---------------------- xmm3 - normalized light vector --
-;---------------------- xmm4 - lo -> hi z1, z2, z3 coords
-;----------------------  as dwords floats ---------------
-;---------------------- xmm5 - lo -> hi y_min, y_max, ---
-;----------------------  x_min, x_max as dword integers -
-;---------------------- stack - no parameters -----------
-;--------------------------------------------------------
-;----------------- procedure don't save registers !! ----
-  push  ebp
-  mov   ebp,esp
-  sub   esp,512
-  sub   ebp,16
-  and   ebp,0xfffffff0
-  .1_nv equ [ebp-16]
-  .2_nv equ [ebp-32]
-  .3_nv equ [ebp-48]
-  .l_v  equ [ebp-64]
-  .z3   equ [ebp-72]
-  .z2   equ [ebp-76]
-  .z1   equ [ebp-80]
-  .x1   equ [ebp-82]
-  .y1   equ [ebp-84]
-  .x2   equ [ebp-86]
-  .y2   equ [ebp-88]
-  .x3   equ [ebp-90]
-  .y3   equ [ebp-92]
-  .Zbuf equ [ebp-96]
-  .x_max equ  [ebp-100]
-  .x_min equ  [ebp-104]
-  .y_max equ  [ebp-108]
-  .y_min equ  [ebp-112]
-  .screen equ [ebp-116]
-  .dx12   equ [ebp-120]
-  .dx13   equ [ebp-124]
-  .dx23   equ [ebp-128]
-  .dn12   equ [ebp-144]
-  .dn13   equ [ebp-160]
-  .dn23   equ [ebp-176]
-  .dz12   equ [ebp-180]
-  .dz13   equ [ebp-184]
-  .dz23   equ [ebp-188]
-  .cnv1   equ [ebp-208]  ; cur normal vectors
-  .cnv2   equ [ebp-224]
-  .cz2    equ [ebp-228]
-  .cz1    equ [ebp-232]
- .sort3:                  ; sort triangle coordinates...
-       cmp     ax,bx
-       jle     .sort1
-       xchg    eax,ebx
-       shufps  xmm4,xmm4,11100001b
-       movaps  xmm6,xmm0
-       movaps  xmm0,xmm1
-       movaps  xmm1,xmm6
- .sort1:
-       cmp      bx,cx
-       jle      .sort2
-       xchg     ebx,ecx
-       shufps   xmm4,xmm4,11011000b
-       movaps   xmm6,xmm1
-       movaps   xmm1,xmm2
-       movaps   xmm2,xmm6
-       jmp .sort3
- .sort2:
-   movaps .z1,xmm4
-   mov    .y1,eax
-   mov    .y2,ebx
-   mov    .y3,ecx
-   movdqa   .y_min,xmm5
-if 1                            ; check if at last only fragment
-   packssdw xmm5,xmm5       ; of triangle is in visable area
-   pshuflw  xmm5,xmm5,11011000b
-   movdqu   xmm7,.y3
-   movdqa   xmm6,xmm5
-   pshufd   xmm5,xmm5,0  ; xmm5 lo-hi -> broadcasted y_min, x_min
-   pshufd   xmm6,xmm6,01010101b ;xmm6 -> brd y_max x_max
-   movdqa   xmm4,xmm7
-   pcmpgtw  xmm7,xmm5
-   pcmpgtw  xmm4,xmm6
-   pxor     xmm7,xmm4
-   pmovmskb eax,xmm7
-   and      eax,0x00aaaaaa
-   or       eax,eax
-   jz       .rpt_loop2_end
-end if
-   movaps   .1_nv,xmm0
-   movaps   .2_nv,xmm1
-   movaps   .3_nv,xmm2
-   movaps   .l_v,xmm3
-   mov      .Zbuf,esi
-   mov      .screen,edi
-       mov      bx,.y2       ; calc deltas
-       sub      bx,.y1
-       jnz      .rpt_dx12_make
-       xorps    xmm7,xmm7
-       mov      dword .dx12,0
-       mov      dword .dz12,0
-       movaps   .dn12,xmm7
-       jmp      .rpt_dx12_done
-  .rpt_dx12_make:
-       mov      ax,.x2
-       sub      ax,.x1
-       cwde
-       movsx    ebx,bx
-       shl      eax,ROUND2
-       cdq
-       idiv     ebx
-       mov      .dx12,eax
-       cvtsi2ss xmm6,ebx
-       movss    xmm5,.z2
-       subss    xmm5,.z1
-       divss    xmm5,xmm6
-       movss    .dz12,xmm5
-       movaps   xmm0,.2_nv
-       subps    xmm0,.1_nv
-       shufps   xmm6,xmm6,0
-       divps    xmm0,xmm6
-       movaps   .dn12,xmm0
-   .rpt_dx12_done:
-       mov      bx,.y3       ; calc deltas
-       sub      bx,.y1
-       jnz      .rpt_dx13_make
-       xorps    xmm7,xmm7
-       mov      dword .dx13,0
-       mov      dword .dz13,0
-       movaps   .dn13,xmm7
-       jmp      .rpt_dx13_done
-  .rpt_dx13_make:
-       mov      ax,.x3
-       sub      ax,.x1
-       cwde
-       movsx    ebx,bx
-       shl      eax,ROUND2
-       cdq
-       idiv     ebx
-       mov      .dx13,eax
-       cvtsi2ss xmm6,ebx
-       movss    xmm5,.z3
-       subss    xmm5,.z1
-       divss    xmm5,xmm6
-       movss    .dz13,xmm5
-       movaps   xmm0,.3_nv
-       subps    xmm0,.1_nv
-       shufps   xmm6,xmm6,0
-       divps    xmm0,xmm6
-       movaps   .dn13,xmm0
-   .rpt_dx13_done:
-       mov      bx,.y3       ; calc deltas
-       sub      bx,.y2
-       jnz      .rpt_dx23_make
-       xorps    xmm7,xmm7
-       mov      dword .dx23,0
-       mov      dword .dz23,0
-       movaps   .dn23,xmm7
-       jmp      .rpt_dx23_done
-  .rpt_dx23_make:
-       mov      ax,.x3
-       sub      ax,.x2
-       cwde
-       movsx    ebx,bx
-       shl      eax,ROUND2
-       cdq
-       idiv     ebx
-       mov      .dx23,eax
-       cvtsi2ss xmm6,ebx
-       movss    xmm5,.z3
-       subss    xmm5,.z2
-       divss    xmm5,xmm6
-       movss    .dz23,xmm5
-       movaps   xmm0,.3_nv
-       subps    xmm0,.2_nv
-       shufps   xmm6,xmm6,0
-       divps    xmm0,xmm6
-       movaps   .dn23,xmm0
-   .rpt_dx23_done:
-       movsx   eax,word .x1
-       shl     eax,ROUND2
-       mov     ebx,eax
-       mov     edx,.z1
-       mov     .cz1,edx
-       mov     .cz2,edx
-       movaps  xmm0,.1_nv
-       movaps  .cnv1,xmm0
-       movaps  .cnv2,xmm0
-       movsx    ecx,word .y1
-       cmp      cx,.y2
-       jge      .rpt_loop1_end
-    .rpt_loop1:
-       pushad
-       movaps   xmm2,.y_min
-       movaps   xmm0,.cnv1
-       movaps   xmm1,.cnv2
-       movlps   xmm3,.cz1
-       movaps   xmm4,.l_v
-       sar      ebx,ROUND2
-       sar      eax,ROUND2
-       mov      edi,.screen
-       mov      esi,.Zbuf
-       call     real_phong_line_z
-       popad
-       movaps   xmm0,.cnv1
-       movaps   xmm1,.cnv2
-       movss    xmm2,.cz1
-       movss    xmm3,.cz2
-       addps    xmm0,.dn13
-       addps    xmm1,.dn12
-       addss    xmm2,.dz13
-       addss    xmm3,.dz12
-       add      eax,.dx13
-       add      ebx,.dx12
-       movaps   .cnv1,xmm0
-       movaps   .cnv2,xmm1
-       movss    .cz1,xmm2
-       movss    .cz2,xmm3
-       add      ecx,1
-       cmp      cx,.y2
-       jl       .rpt_loop1
-   .rpt_loop1_end:
-       movsx    ecx,word .y2
-       cmp      cx,.y3
-       jge      .rpt_loop2_end
-       movsx    ebx,word .x2                    ; eax - cur x1
-       shl      ebx,ROUND2                 ; ebx - cur x2
-       push     dword .z2
-       pop      dword .cz2
-       movaps   xmm0,.2_nv
-       movaps   .cnv2,xmm0
-     .rpt_loop2:
-       pushad
-       movaps   xmm2,.y_min
-       movaps   xmm0,.cnv1
-       movaps   xmm1,.cnv2
-       movlps   xmm3,.cz1
-       movaps   xmm4,.l_v
-       sar      ebx,ROUND2
-       sar      eax,ROUND2
-       mov      edi,.screen
-       mov      esi,.Zbuf
-       call     real_phong_line_z
-       popad
-       movaps   xmm0,.cnv1
-       movaps   xmm1,.cnv2
-       movss    xmm2,.cz1
-       movss    xmm3,.cz2
-       addps    xmm0,.dn13
-       addps    xmm1,.dn23
-       addss    xmm2,.dz13
-       addss    xmm3,.dz23
-       add      eax,.dx13
-       add      ebx,.dx23
-       movaps   .cnv1,xmm0
-       movaps   .cnv2,xmm1
-       movss    .cz1,xmm2
-       movss    .cz2,xmm3
-       add      ecx,1
-       cmp      cx,.y3
-       jl       .rpt_loop2
-    .rpt_loop2_end:
-      add   esp,512
-      pop   ebp
-ret
-align 16
-real_phong_line_z:
-; in:
-;    xmm0 - normal vector 1
-;    xmm1 - normal vect 2
-;    xmm3 - lo -> hi z1, z2 coords as dwords floats
-;    xmm2 - lo -> hi y_min, y_max, x_min, x_max
-;           as dword integers
-;    xmm4 - normalized light vector
-;    eax - x1
-;    ebx - x2
-;    ecx - y
-;    edi - screen buffer
-;    esi - z buffer filled with dd floats
-   push  ebp
-   mov   ebp,esp
-   sub   esp,160
-   sub   ebp,16
-   and   ebp,0xfffffff0
- .n1 equ [ebp-16]
- .n2 equ [ebp-32]
- .lv equ [ebp-48]
- .lx1 equ [ebp-52]
- .lx2 equ [ebp-56]
- .z2 equ [ebp-60]
- .z1 equ [ebp-64]
- .screen equ [ebp-68]
- .zbuff  equ [ebp-72]
- .x_max  equ [ebp-74]
- .x_min  equ [ebp-76]
- .y_max  equ [ebp-78]
- .y_min  equ [ebp-80]
- .dn     equ [ebp-96]
- .dz     equ [ebp-100]
- .y      equ [ebp-104]
- .cnv    equ [ebp-128]
-        mov    .y,ecx
-        packssdw xmm2,xmm2
-        movq   .y_min,xmm2
-        cmp    cx,.y_min
-        jl     .end_rp_line
-        cmp    cx,.y_max
-        jge    .end_rp_line          ;
-        cmp     eax,ebx
-        je      .end_rp_line
-        jl      @f
-        xchg    eax,ebx
-        movaps  xmm7,xmm0
-        movaps  xmm0,xmm1
-        movaps  xmm1,xmm7
-        shufps  xmm3,xmm3,11100001b
-   @@:
-        cmp     ax,.x_max
-        jge     .end_rp_line
-        cmp     bx,.x_min
-        jle     .end_rp_line
-        movaps  .lv,xmm4
-        movaps  .n1,xmm0
-        movaps  .n2,xmm1
-        mov     .lx1,eax
-        mov     .lx2,ebx
-        movlps  .z1,xmm3
-        sub     ebx,eax
-        cvtsi2ss xmm7,ebx
-        shufps  xmm7,xmm7,0
-        subps   xmm1,xmm0
-        divps   xmm1,xmm7
-        movaps  .dn,xmm1
-        psrldq  xmm3,4
-        subss   xmm3,.z1
-        divss   xmm3,xmm7
-        movss   .dz,xmm3
-        mov      ebx,.lx1
-        cmp      bx,.x_min     ; clipping on function4
-        jge      @f
-        movzx    eax,word .x_min
-        sub      eax,ebx
-        cvtsi2ss xmm7,eax
-        shufps   xmm7,xmm7,0
-        mulss    xmm3,xmm7
-        mulps    xmm1,xmm7
-        addss    xmm3,.z1
-        addps    xmm1,.n1
-        movsx    eax,word .x_min
-        movss    .z1,xmm3
-        movaps   .n1,xmm1
-        mov      dword .lx1,eax
-      @@:
-        movzx   eax,word .x_max
-        cmp     .lx2,eax
-        jl      @f
-        mov     .lx2,eax
-      @@:
-        movzx   eax,word[size_x_var]
-        mul     dword .y
-      ;  mov     edx,.x1
-        add     eax,.lx1
-        shl     eax,2
-        add     edi,eax
-        add     esi,eax
-        mov     ecx,.lx2
-        sub     ecx,.lx1
-        movaps  xmm0,.n1
-        movss   xmm2,.z1
-align 16
-   .ddraw:
-        movss    xmm7,xmm2
-        cmpnltss xmm7,dword[esi]
-        movd     eax,xmm7
-        or       eax,eax
-        jnz      .skip
-        movss    [esi],xmm2
-        movaps   xmm7,xmm0
-        mulps    xmm7,xmm7 ; normalize
-        haddps   xmm7,xmm7
-        haddps   xmm7,xmm7
-        rsqrtps  xmm7,xmm7
-        mulps    xmm7,xmm0
-        movaps   .cnv,xmm7
-        mov      edx,lights_aligned    ; lights - global variable
-        xorps    xmm1,xmm1     ; instead global can be used .lv - light vect.
-      @@:
-        movaps    xmm6,[edx+16]
-        movaps    xmm5,[edx]
-        movaps    xmm3,[edx+48]
-        andps     xmm5,[zero_hgst_dd]  ; global
-        mulps    xmm5,.cnv  ;.lv  ; last dword should be zeroed
-        haddps   xmm5,xmm5
-        haddps   xmm5,xmm5
-   ;     mulps    xmm5,[env_const2]
-   ;     maxps    xmm5,[dot_min]
-   ;     minps    xmm5,[dot_max]
-        movaps   xmm7,xmm5
-      ;  mulps    xmm7,[env_const2]
-     ;   mulps    xmm7,[env_const2]
-     ;   maxps    xmm7,[dot_min]
-     ;   minps    xmm7,[dot_max]
-        mulps    xmm7,xmm7
-        mulps    xmm7,xmm7
-        mulps    xmm5,xmm6
-        mulps    xmm7,xmm7
-        mulps    xmm7,xmm3
-        addps    xmm5,xmm7
-        minps    xmm5,[mask_255f]   ; global
-        maxps    xmm1,xmm5
-     ;   movq     xmm3,[edx+20]    ; minimal color
-     ;   punpcklwd xmm3,[minimum0]
-     ;   cvtdq2ps xmm3,xmm3
-     ;   maxps    xmm1,xmm3
-        add      edx,64
-        cmp      edx,lights_aligned_end    ; global
-        jnz      @b
-        cvtps2dq xmm1,xmm1
-        packssdw xmm1,xmm1
-        packuswb xmm1,xmm1
-        movd     [edi],xmm1
-     .skip:
-        add      edi,4
-        add      esi,4
-        addps    xmm0,.dn
-        addss    xmm2,.dz
-        sub      ecx,1
-        jnz      .ddraw
-  .end_rp_line:
-        add      esp,160
-        pop      ebp
-ret
+; Real Phong's shading implemented if flat assembler
+; by Maciej Guba.
+; http://macgub.co.pl
+ROUND2 equ 10
+real_phong_tri_z:
+;----procedure render Phongs shaded triangle with z coord
+;----interpolation ( Catmull alghoritm )-----------------
+;----I normalize normal vector in every pixel -----------
+;------------------in - eax - x1 shl 16 + y1 ------------
+;---------------------- ebx - x2 shl 16 + y2 ------------
+;---------------------- ecx - x3 shl 16 + y3 ------------
+;---------------------- esi - pointer to Z-buffer filled-
+;----------------------   with dd float variables--------
+;---------------------- edi - pointer to screen buffer---
+;---------------------- xmm0 - 1st normal vector --------
+;---------------------- xmm1 - 2cond normal vector ------
+;---------------------- xmm2 - 3rd normal vector --------
+;---------------------- xmm3 - normalized light vector --
+;---------------------- xmm4 - lo -> hi z1, z2, z3 coords
+;----------------------  as dwords floats ---------------
+;---------------------- xmm5 - lo -> hi y_min, y_max, ---
+;----------------------  x_min, x_max as dword integers -
+;---------------------- stack - no parameters -----------
+;--------------------------------------------------------
+;----------------- procedure don't save registers !! ----
+  push  ebp
+  mov   ebp,esp
+  sub   esp,512
+  sub   ebp,16
+  and   ebp,0xfffffff0
+  .1_nv equ [ebp-16]
+  .2_nv equ [ebp-32]
+  .3_nv equ [ebp-48]
+  .l_v  equ [ebp-64]
+  .z3   equ [ebp-72]
+  .z2   equ [ebp-76]
+  .z1   equ [ebp-80]
+  .x1   equ [ebp-82]
+  .y1   equ [ebp-84]
+  .x2   equ [ebp-86]
+  .y2   equ [ebp-88]
+  .x3   equ [ebp-90]
+  .y3   equ [ebp-92]
+  .Zbuf equ [ebp-96]
+  .x_max equ  [ebp-100]
+  .x_min equ  [ebp-104]
+  .y_max equ  [ebp-108]
+  .y_min equ  [ebp-112]
+  .screen equ [ebp-116]
+  .dx12   equ [ebp-120]
+  .dx13   equ [ebp-124]
+  .dx23   equ [ebp-128]
+  .dn12   equ [ebp-144]
+  .dn13   equ [ebp-160]
+  .dn23   equ [ebp-176]
+  .dz12   equ [ebp-180]
+  .dz13   equ [ebp-184]
+  .dz23   equ [ebp-188]
+  .cnv1   equ [ebp-208]  ; cur normal vectors
+  .cnv2   equ [ebp-224]
+  .cz2    equ [ebp-228]
+  .cz1    equ [ebp-232]
+ .sort3:                  ; sort triangle coordinates...
+       cmp     ax,bx
+       jle     .sort1
+       xchg    eax,ebx
+       shufps  xmm4,xmm4,11100001b
+       movaps  xmm6,xmm0
+       movaps  xmm0,xmm1
+       movaps  xmm1,xmm6
+ .sort1:
+       cmp      bx,cx
+       jle      .sort2
+       xchg     ebx,ecx
+       shufps   xmm4,xmm4,11011000b
+       movaps   xmm6,xmm1
+       movaps   xmm1,xmm2
+       movaps   xmm2,xmm6
+       jmp .sort3
+ .sort2:
+   movaps .z1,xmm4
+   mov    .y1,eax
+   mov    .y2,ebx
+   mov    .y3,ecx
+   movdqa   .y_min,xmm5
+if 1                            ; check if at last only fragment
+   packssdw xmm5,xmm5       ; of triangle is in visable area
+   pshuflw  xmm5,xmm5,11011000b
+   movdqu   xmm7,.y3
+   movdqa   xmm6,xmm5
+   pshufd   xmm5,xmm5,0  ; xmm5 lo-hi -> broadcasted y_min, x_min
+   pshufd   xmm6,xmm6,01010101b ;xmm6 -> brd y_max x_max
+   movdqa   xmm4,xmm7
+   pcmpgtw  xmm7,xmm5
+   pcmpgtw  xmm4,xmm6
+   pxor     xmm7,xmm4
+   pmovmskb eax,xmm7
+   and      eax,0x00aaaaaa
+   or       eax,eax
+   jz       .rpt_loop2_end
+end if
+   movaps   .1_nv,xmm0
+   movaps   .2_nv,xmm1
+   movaps   .3_nv,xmm2
+   movaps   .l_v,xmm3
+   mov      .Zbuf,esi
+   mov      .screen,edi
+       mov      bx,.y2       ; calc deltas
+       sub      bx,.y1
+       jnz      .rpt_dx12_make
+       xorps    xmm7,xmm7
+       mov      dword .dx12,0
+       mov      dword .dz12,0
+       movaps   .dn12,xmm7
+       jmp      .rpt_dx12_done
+  .rpt_dx12_make:
+       mov      ax,.x2
+       sub      ax,.x1
+       cwde
+       movsx    ebx,bx
+       shl      eax,ROUND2
+       cdq
+       idiv     ebx
+       mov      .dx12,eax
+       cvtsi2ss xmm6,ebx
+       movss    xmm5,.z2
+       subss    xmm5,.z1
+       divss    xmm5,xmm6
+       movss    .dz12,xmm5
+       movaps   xmm0,.2_nv
+       subps    xmm0,.1_nv
+       shufps   xmm6,xmm6,0
+       divps    xmm0,xmm6
+       movaps   .dn12,xmm0
+   .rpt_dx12_done:
+       mov      bx,.y3       ; calc deltas
+       sub      bx,.y1
+       jnz      .rpt_dx13_make
+       xorps    xmm7,xmm7
+       mov      dword .dx13,0
+       mov      dword .dz13,0
+       movaps   .dn13,xmm7
+       jmp      .rpt_dx13_done
+  .rpt_dx13_make:
+       mov      ax,.x3
+       sub      ax,.x1
+       cwde
+       movsx    ebx,bx
+       shl      eax,ROUND2
+       cdq
+       idiv     ebx
+       mov      .dx13,eax
+       cvtsi2ss xmm6,ebx
+       movss    xmm5,.z3
+       subss    xmm5,.z1
+       divss    xmm5,xmm6
+       movss    .dz13,xmm5
+       movaps   xmm0,.3_nv
+       subps    xmm0,.1_nv
+       shufps   xmm6,xmm6,0
+       divps    xmm0,xmm6
+       movaps   .dn13,xmm0
+   .rpt_dx13_done:
+       mov      bx,.y3       ; calc deltas
+       sub      bx,.y2
+       jnz      .rpt_dx23_make
+       xorps    xmm7,xmm7
+       mov      dword .dx23,0
+       mov      dword .dz23,0
+       movaps   .dn23,xmm7
+       jmp      .rpt_dx23_done
+  .rpt_dx23_make:
+       mov      ax,.x3
+       sub      ax,.x2
+       cwde
+       movsx    ebx,bx
+       shl      eax,ROUND2
+       cdq
+       idiv     ebx
+       mov      .dx23,eax
+       cvtsi2ss xmm6,ebx
+       movss    xmm5,.z3
+       subss    xmm5,.z2
+       divss    xmm5,xmm6
+       movss    .dz23,xmm5
+       movaps   xmm0,.3_nv
+       subps    xmm0,.2_nv
+       shufps   xmm6,xmm6,0
+       divps    xmm0,xmm6
+       movaps   .dn23,xmm0
+   .rpt_dx23_done:
+       movsx   eax,word .x1
+       shl     eax,ROUND2
+       mov     ebx,eax
+       mov     edx,.z1
+       mov     .cz1,edx
+       mov     .cz2,edx
+       movaps  xmm0,.1_nv
+       movaps  .cnv1,xmm0
+       movaps  .cnv2,xmm0
+       movsx    ecx,word .y1
+       cmp      cx,.y2
+       jge      .rpt_loop1_end
+    .rpt_loop1:
+       pushad
+       movaps   xmm2,.y_min
+       movaps   xmm0,.cnv1
+       movaps   xmm1,.cnv2
+       movlps   xmm3,.cz1
+       movaps   xmm4,.l_v
+       sar      ebx,ROUND2
+       sar      eax,ROUND2
+       mov      edi,.screen
+       mov      esi,.Zbuf
+       call     real_phong_line_z
+       popad
+       movaps   xmm0,.cnv1
+       movaps   xmm1,.cnv2
+       movss    xmm2,.cz1
+       movss    xmm3,.cz2
+       addps    xmm0,.dn13
+       addps    xmm1,.dn12
+       addss    xmm2,.dz13
+       addss    xmm3,.dz12
+       add      eax,.dx13
+       add      ebx,.dx12
+       movaps   .cnv1,xmm0
+       movaps   .cnv2,xmm1
+       movss    .cz1,xmm2
+       movss    .cz2,xmm3
+       add      ecx,1
+       cmp      cx,.y2
+       jl       .rpt_loop1
+   .rpt_loop1_end:
+       movsx    ecx,word .y2
+       cmp      cx,.y3
+       jge      .rpt_loop2_end
+       movsx    ebx,word .x2                    ; eax - cur x1
+       shl      ebx,ROUND2                 ; ebx - cur x2
+       push     dword .z2
+       pop      dword .cz2
+       movaps   xmm0,.2_nv
+       movaps   .cnv2,xmm0
+     .rpt_loop2:
+       pushad
+       movaps   xmm2,.y_min
+       movaps   xmm0,.cnv1
+       movaps   xmm1,.cnv2
+       movlps   xmm3,.cz1
+       movaps   xmm4,.l_v
+       sar      ebx,ROUND2
+       sar      eax,ROUND2
+       mov      edi,.screen
+       mov      esi,.Zbuf
+       call     real_phong_line_z
+       popad
+       movaps   xmm0,.cnv1
+       movaps   xmm1,.cnv2
+       movss    xmm2,.cz1
+       movss    xmm3,.cz2
+       addps    xmm0,.dn13
+       addps    xmm1,.dn23
+       addss    xmm2,.dz13
+       addss    xmm3,.dz23
+       add      eax,.dx13
+       add      ebx,.dx23
+       movaps   .cnv1,xmm0
+       movaps   .cnv2,xmm1
+       movss    .cz1,xmm2
+       movss    .cz2,xmm3
+       add      ecx,1
+       cmp      cx,.y3
+       jl       .rpt_loop2
+    .rpt_loop2_end:
+      add   esp,512
+      pop   ebp
+ret
+align 16
+real_phong_line_z:
+; in:
+;    xmm0 - normal vector 1
+;    xmm1 - normal vect 2
+;    xmm3 - lo -> hi z1, z2 coords as dwords floats
+;    xmm2 - lo -> hi y_min, y_max, x_min, x_max
+;           as dword integers
+;    xmm4 - normalized light vector
+;    eax - x1
+;    ebx - x2
+;    ecx - y
+;    edi - screen buffer
+;    esi - z buffer filled with dd floats
+   push  ebp
+   mov   ebp,esp
+   sub   esp,160
+   sub   ebp,16
+   and   ebp,0xfffffff0
+ .n1 equ [ebp-16]
+ .n2 equ [ebp-32]
+ .lv equ [ebp-48]
+ .lx1 equ [ebp-52]
+ .lx2 equ [ebp-56]
+ .z2 equ [ebp-60]
+ .z1 equ [ebp-64]
+ .screen equ [ebp-68]
+ .zbuff  equ [ebp-72]
+ .x_max  equ [ebp-74]
+ .x_min  equ [ebp-76]
+ .y_max  equ [ebp-78]
+ .y_min  equ [ebp-80]
+ .dn     equ [ebp-96]
+ .dz     equ [ebp-100]
+ .y      equ [ebp-104]
+ .cnv    equ [ebp-128]
+        mov    .y,ecx
+        packssdw xmm2,xmm2
+        movq   .y_min,xmm2
+        cmp    cx,.y_min
+        jl     .end_rp_line
+        cmp    cx,.y_max
+        jge    .end_rp_line          ;
+        cmp     eax,ebx
+        je      .end_rp_line
+        jl      @f
+        xchg    eax,ebx
+        movaps  xmm7,xmm0
+        movaps  xmm0,xmm1
+        movaps  xmm1,xmm7
+        shufps  xmm3,xmm3,11100001b
+   @@:
+        cmp     ax,.x_max
+        jge     .end_rp_line
+        cmp     bx,.x_min
+        jle     .end_rp_line
+        movaps  .lv,xmm4
+        movaps  .n1,xmm0
+        movaps  .n2,xmm1
+        mov     .lx1,eax
+        mov     .lx2,ebx
+        movlps  .z1,xmm3
+        sub     ebx,eax
+        cvtsi2ss xmm7,ebx
+        shufps  xmm7,xmm7,0
+        subps   xmm1,xmm0
+        divps   xmm1,xmm7
+        movaps  .dn,xmm1
+        psrldq  xmm3,4
+        subss   xmm3,.z1
+        divss   xmm3,xmm7
+        movss   .dz,xmm3
+        mov      ebx,.lx1
+        cmp      bx,.x_min     ; clipping on function4
+        jge      @f
+        movzx    eax,word .x_min
+        sub      eax,ebx
+        cvtsi2ss xmm7,eax
+        shufps   xmm7,xmm7,0
+        mulss    xmm3,xmm7
+        mulps    xmm1,xmm7
+        addss    xmm3,.z1
+        addps    xmm1,.n1
+        movsx    eax,word .x_min
+        movss    .z1,xmm3
+        movaps   .n1,xmm1
+        mov      dword .lx1,eax
+      @@:
+        movzx   eax,word .x_max
+        cmp     .lx2,eax
+        jl      @f
+        mov     .lx2,eax
+      @@:
+        movzx   eax,word[size_x_var]
+        mul     dword .y
+      ;  mov     edx,.x1
+        add     eax,.lx1
+        shl     eax,2
+        add     edi,eax
+        add     esi,eax
+        mov     ecx,.lx2
+        sub     ecx,.lx1
+        movaps  xmm0,.n1
+        movss   xmm2,.z1
+align 16
+   .ddraw:
+        movss    xmm7,xmm2
+        cmpnltss xmm7,dword[esi]
+        movd     eax,xmm7
+        or       eax,eax
+        jnz      .skip
+        movss    [esi],xmm2
+        movaps   xmm7,xmm0
+        mulps    xmm7,xmm7 ; normalize
+        haddps   xmm7,xmm7
+        haddps   xmm7,xmm7
+        rsqrtps  xmm7,xmm7
+        mulps    xmm7,xmm0
+        movaps   .cnv,xmm7
+        mov      edx,lights_aligned    ; lights - global variable
+        xorps    xmm1,xmm1     ; instead global can be used .lv - light vect.
+      @@:
+        movaps    xmm6,[edx+16]
+        movaps    xmm5,[edx]
+        movaps    xmm3,[edx+48]
+        andps     xmm5,[zero_hgst_dd]  ; global
+        mulps    xmm5,.cnv  ;.lv  ; last dword should be zeroed
+        haddps   xmm5,xmm5
+        haddps   xmm5,xmm5
+   ;     mulps    xmm5,[env_const2]
+   ;     maxps    xmm5,[dot_min]
+   ;     minps    xmm5,[dot_max]
+        movaps   xmm7,xmm5
+      ;  mulps    xmm7,[env_const2]
+     ;   mulps    xmm7,[env_const2]
+     ;   maxps    xmm7,[dot_min]
+     ;   minps    xmm7,[dot_max]
+        mulps    xmm7,xmm7
+        mulps    xmm7,xmm7
+        mulps    xmm5,xmm6
+        mulps    xmm7,xmm7
+        mulps    xmm7,xmm3
+        addps    xmm5,xmm7
+        minps    xmm5,[mask_255f]   ; global
+        maxps    xmm1,xmm5
+     ;   movq     xmm3,[edx+20]    ; minimal color
+     ;   punpcklwd xmm3,[minimum0]
+     ;   cvtdq2ps xmm3,xmm3
+     ;   maxps    xmm1,xmm3
+        add      edx,64
+        cmp      edx,lights_aligned_end    ; global
+        jnz      @b
+        cvtps2dq xmm1,xmm1
+        packssdw xmm1,xmm1
+        packuswb xmm1,xmm1
+        movd     [edi],xmm1
+     .skip:
+        add      edi,4
+        add      esi,4
+        addps    xmm0,.dn
+        addss    xmm2,.dz
+        sub      ecx,1
+        jnz      .ddraw
+  .end_rp_line:
+        add      esp,160
+        pop      ebp
+ret

 /programs/demos/view3ds/3ray_shd.inc
 ,0 → 1,688
+; Ray casted shadows
+; by Maciej Guba.
+; http://macgub.co.pl
+ROUND2 equ 10
+ray_shad:
+;--- Procedure render triangle with ray casted shadow ---
+;--- effect. Calc intersection with all triangles in ----
+;--- everypixel. Its not real time process, especially --
+;--- when many triangles are computed. ------------------
+;------in - eax - x1 shl 16 + y1 ------------------------
+;---------- ebx - x2 shl 16 + y2 ------------------------
+;---------- ecx - x3 shl 16 + y3 ------------------------
+;---------- edx - ptr to fur coords struct --------------
+;---------- esi - pointer to stencil / Z-buffer, filled -
+;--------------   with dword float variables, it masks --
+;--------------   'Z' position (coord) of every front ---
+;--------------   pixel. --------------------------------
+;---------- edi - pointer to screen buffer --------------
+;---------- xmm0 - 1st normal vector --------------------
+;---------- xmm1 - 2cond normal vector ------------------
+;---------- xmm2 - 3rd normal vector --------------------
+;---------- xmm3 - --------------------------------------
+;---------- xmm4 - lo -> hi z1, z2, z3 coords -----------
+;---------------   as dwords floats ---------------------
+;---------- xmm5 - lo -> hi y_min, y_max, x_min, x_max --
+;---------------   as dword integers --------------------
+;-----------mm7  - current triangle index ---------------
+;---------------------- stack - no parameters -----------
+;--------------------------------------------------------
+;----------------- procedure don't save registers !! ----
+  push  ebp
+  mov   ebp,esp
+  sub   esp,1024
+  sub   ebp,16
+  and   ebp,0xfffffff0
+  .1_nv equ [ebp-16]
+  .2_nv equ [ebp-32]
+  .3_nv equ [ebp-48]
+  .l_v  equ [ebp-64]
+  .z3   equ [ebp-72]
+  .z2   equ [ebp-76]
+  .z1   equ [ebp-80]
+  .x1   equ [ebp-82]
+  .y1   equ [ebp-84]
+  .x2   equ [ebp-86]
+  .y2   equ [ebp-88]
+  .x3   equ [ebp-90]
+  .y3   equ [ebp-92]
+  .Zbuf equ [ebp-96]
+  .x_max  equ [ebp-100]
+  .x_min  equ [ebp-104]
+  .y_max  equ [ebp-108]
+  .y_min  equ [ebp-112]
+  .screen equ [ebp-116]
+  .dx12   equ [ebp-120]
+  .dx13   equ [ebp-124]
+  .dx23   equ [ebp-128]
+  .dn12   equ [ebp-144]
+  .dn13   equ [ebp-160]
+  .dn23   equ [ebp-176]
+  .dz12   equ [ebp-180]
+  .dz13   equ [ebp-184]
+  .dz23   equ [ebp-188]
+  .cnv1   equ [ebp-208] ; current normal vectors
+  .cnv2   equ [ebp-240]
+  .cz2    equ [ebp-244]
+  .cz1    equ [ebp-248]
+  .tri_no equ [ebp-252]
+ .sort3:                  ; sort triangle coordinates...
+       cmp     ax,bx
+       jle     .sort1
+       xchg    eax,ebx
+       shufps  xmm4,xmm4,11100001b
+       movaps  xmm6,xmm0
+       movaps  xmm0,xmm1
+       movaps  xmm1,xmm6
+ .sort1:
+       cmp      bx,cx
+       jle      .sort2
+       xchg     ebx,ecx
+       shufps   xmm4,xmm4,11011000b
+       movaps   xmm6,xmm1
+       movaps   xmm1,xmm2
+       movaps   xmm2,xmm6
+       jmp .sort3
+ .sort2:
+   movaps .z1,xmm4
+   mov    .y1,eax
+   mov    .y2,ebx
+   mov    .y3,ecx
+   movdqa   .y_min,xmm5
+if 1                            ; check if at last only fragment
+   packssdw xmm5,xmm5       ; of triangle is in visable area
+   pshuflw  xmm5,xmm5,11011000b
+   movdqu   xmm7,.y3
+   movdqa   xmm6,xmm5
+   pshufd   xmm5,xmm5,0  ; xmm5 lo-hi -> broadcasted y_min, x_min
+   pshufd   xmm6,xmm6,01010101b ;xmm6 -> brd y_max x_max
+   movdqa   xmm4,xmm7
+   pcmpgtw  xmm7,xmm5
+   pcmpgtw  xmm4,xmm6
+   pxor     xmm7,xmm4
+   pmovmskb eax,xmm7
+   and      eax,0x00aaaaaa
+   or       eax,eax
+   jz       .rpt_loop2_end
+end if
+   movd     .tri_no,mm7
+   movaps   .1_nv,xmm0
+   movaps   .2_nv,xmm1
+   movaps   .3_nv,xmm2
+ ;  movaps   .l_v,xmm3
+   mov      .Zbuf,esi
+   mov      .screen,edi
+       mov      bx,.y2       ; calc deltas
+       sub      bx,.y1
+       jnz      .rpt_dx12_make
+       xorps    xmm7,xmm7
+       mov      dword .dx12,0
+       mov      dword .dz12,0
+       movaps   .dn12,xmm7
+       jmp      .rpt_dx12_done
+  .rpt_dx12_make:
+       mov      ax,.x2
+       sub      ax,.x1
+       cwde
+       movsx    ebx,bx
+       shl      eax,ROUND2
+       cdq
+       idiv     ebx
+       mov      .dx12,eax
+       cvtsi2ss xmm6,ebx
+       movss    xmm5,.z2
+       rcpss    xmm6,xmm6
+       subss    xmm5,.z1
+       mulss    xmm5,xmm6
+       movss    .dz12,xmm5
+       shufps   xmm6,xmm6,0
+       movaps   xmm0,.2_nv
+       subps    xmm0,.1_nv
+       mulps    xmm0,xmm6
+       movaps   .dn12,xmm0
+;       subps    xmm3,xmm0
+;       mulps    xmm3,xmm6
+   .rpt_dx12_done:
+       mov      bx,.y3       ; calc deltas
+       sub      bx,.y1
+       jnz      .rpt_dx13_make
+       xorps    xmm7,xmm7
+       mov      dword .dx13,0
+       mov      dword .dz13,0
+       movaps   .dn13,xmm7
+       jmp      .rpt_dx13_done
+  .rpt_dx13_make:
+       mov      ax,.x3
+       sub      ax,.x1
+       cwde
+       movsx    ebx,bx
+       shl      eax,ROUND2
+       cdq
+       idiv     ebx
+       mov      .dx13,eax
+       cvtsi2ss xmm6,ebx
+       movss    xmm5,.z3
+       rcpss    xmm6,xmm6
+       subss    xmm5,.z1
+       mulss    xmm5,xmm6
+       movss    .dz13,xmm5
+       movaps   xmm0,.3_nv
+       subps    xmm0,.1_nv
+       shufps   xmm6,xmm6,0
+       mulps    xmm0,xmm6
+       movaps   .dn13,xmm0
+ ;      mulps    xmm0,xmm6
+  .rpt_dx13_done:
+       mov      bx,.y3       ; calc deltas
+       sub      bx,.y2
+       jnz      .rpt_dx23_make
+       xorps    xmm7,xmm7
+       mov      dword .dx23,0
+       mov      dword .dz23,0
+       movaps   .dn23,xmm7
+       jmp      .rpt_dx23_done
+  .rpt_dx23_make:
+       mov      ax,.x3
+       sub      ax,.x2
+       cwde
+       movsx    ebx,bx
+       shl      eax,ROUND2
+       cdq
+       idiv     ebx
+       mov      .dx23,eax
+       cvtsi2ss xmm6,ebx
+       movss    xmm5,.z3
+       rcpss    xmm6,xmm6
+       subss    xmm5,.z2
+       mulss    xmm5,xmm6
+       movss    .dz23,xmm5
+       movaps   xmm0,.3_nv
+       subps    xmm0,.2_nv
+       shufps   xmm6,xmm6,0
+       mulps    xmm0,xmm6
+       movaps   .dn23,xmm0
+  ;     mulps    xmm0,xmm6
+   .rpt_dx23_done:
+       movsx   eax,word .x1
+       shl     eax,ROUND2
+       mov     ebx,eax
+       mov     ecx,.z1
+       mov     .cz1,ecx
+       mov     .cz2,ecx
+       movaps  xmm0,.1_nv
+       movaps  .cnv1,xmm0
+       movaps  .cnv2,xmm0
+       mov      edi,.screen
+       mov      esi,.Zbuf
+       movsx    ecx,word .y1
+       cmp      cx,.y2
+       jge      .rpt_loop1_end
+    .rpt_loop1:
+       pushad
+       movaps   xmm2,.y_min
+       movaps   xmm0,.cnv1
+       movaps   xmm1,.cnv2
+       movlps   xmm3,.cz1
+    ;   movaps   xmm4,.l_v
+       sar      ebx,ROUND2
+       sar      eax,ROUND2
+       movd     mm7,.tri_no
+       call     ray_shd_l
+       popad
+       movaps   xmm0,.cnv1
+       movaps   xmm1,.cnv2
+       ; fur x,y
+       movss    xmm2,.cz1
+       movss    xmm3,.cz2
+       shufps   xmm4,xmm4,01001110b
+       addps    xmm0,.dn13
+       addps    xmm1,.dn12
+       addss    xmm2,.dz13
+       addss    xmm3,.dz12
+       add      eax,.dx13
+       add      ebx,.dx12
+       shufps   xmm4,xmm4,01001110b
+       movaps   .cnv1,xmm0
+       movaps   .cnv2,xmm1
+       movss    .cz1,xmm2
+       movss    .cz2,xmm3
+       add      ecx,1
+       cmp      cx,.y2
+       jl       .rpt_loop1
+   .rpt_loop1_end:
+       movsx    ecx,word .y2
+       cmp      cx,.y3
+       jge      .rpt_loop2_end
+       movsx    ebx,word .x2               ; eax - cur x1
+       shl      ebx,ROUND2                 ; ebx - cur x2
+       push     dword .z2
+       pop      dword .cz2
+       movaps   xmm0,.2_nv
+       movaps   .cnv2,xmm0
+       mov      edi,.screen
+       mov      esi,.Zbuf
+     .rpt_loop2:
+       pushad
+       movaps   xmm2,.y_min
+       movaps   xmm0,.cnv1
+       movaps   xmm1,.cnv2
+       movlps   xmm3,.cz1
+   ;    movaps   xmm4,.l_v
+       sar      ebx,ROUND2
+       sar      eax,ROUND2
+       movd     mm7,.tri_no
+       call     ray_shd_l
+       popad
+       movaps   xmm0,.cnv1
+       movaps   xmm1,.cnv2
+       movss    xmm2,.cz1
+       movss    xmm3,.cz2
+       addps    xmm0,.dn13
+       addps    xmm1,.dn23
+       addss    xmm2,.dz13
+       addss    xmm3,.dz23
+       add      eax,.dx13
+       add      ebx,.dx23
+       addps    xmm4,xmm6
+       movaps   .cnv1,xmm0
+       movaps   .cnv2,xmm1
+       movss    .cz1,xmm2
+       movss    .cz2,xmm3
+       add      ecx,1
+       cmp      cx,.y3
+       jl       .rpt_loop2
+    .rpt_loop2_end:
+      add   esp,1024
+      pop   ebp
+ret
+align 16
+ray_shd_l:
+; in:
+;    xmm0 - normal vector 1
+;    xmm1 - normal vect 2
+;    xmm3 - lo -> hi z1, z2 coords as dwords floats
+;    xmm2 - lo -> hi y_min, y_max, x_min, x_max
+;           as dword integers
+;    xmm4 - ----
+;    mm7  - current triangle index
+;    eax  - x1
+;    ebx  - x2
+;    ecx  - y
+;    edx  - -----
+;    edi  - screen buffer
+;    esi  - z buffer / stencil buffer filled with dd floats
+   push  ebp
+   mov   ebp,esp
+   sub   esp,320
+   sub   ebp,16
+   and   ebp,0xfffffff0
+ .n1  equ [ebp-16]
+ .n2  equ [ebp-32]
+ .lv  equ [ebp-48]
+ .lx1 equ [ebp-52]
+ .lx2 equ [ebp-56]
+ .z2  equ [ebp-60]
+ .z1  equ [ebp-64]
+ .screen    equ [ebp-68]
+ .zbuff     equ [ebp-72]
+ .x_max     equ [ebp-74]
+ .x_min     equ [ebp-76]
+ .y_max     equ [ebp-78]
+ .y_min     equ [ebp-80]
+ .dn        equ [ebp-96]
+ .dz        equ [ebp-100]
+ .y         equ [ebp-104]
+; .cur_tri   equ [ebp-108]
+ .cnv       equ [ebp-128]
+ .Rlen      equ [ebp-128-16]
+ .r1        equ [ebp-128-32]
+ .vect_t    equ [ebp-128-48]
+ .cur_tri    equ [ebp-128-64]
+; .p3t       equ [ebp-128-80]
+ .nray      equ [ebp-128-96]
+ .final_col equ [ebp-128-112]
+ .aabb_mask equ dword[ebp-128-112-4]
+        mov     .y,ecx
+        movdqa   xmm4,xmm2
+        packssdw xmm2,xmm2
+        movq    .y_min,xmm2
+        cmp     cx,.y_min
+        jl      .end_rp_line
+        cmp     cx,.y_max
+        jge     .end_rp_line          ;
+        cmp     eax,ebx
+        je      .end_rp_line
+        jl      @f
+        xchg    eax,ebx
+        movaps  xmm7,xmm0
+        movaps  xmm0,xmm1
+        movaps  xmm1,xmm7
+        shufps  xmm3,xmm3,11100001b
+    @@:
+        movd    .cur_tri,mm7
+        cmp     ax,.x_max
+        jge     .end_rp_line
+        cmp     bx,.x_min
+        jle     .end_rp_line
+     ;   movaps  .lv,xmm4
+        andps    xmm0,[zero_hgst_dd]
+        andps    xmm1,[zero_hgst_dd]
+        movaps  .n1,xmm0
+        movaps  .n2,xmm1
+        mov     .lx1,eax
+        mov     .lx2,ebx
+        movlps  .z1,xmm3
+        sub     ebx,eax
+        cvtsi2ss xmm7,ebx
+        rcpss   xmm7,xmm7
+        shufps  xmm7,xmm7,0
+        subps   xmm1,xmm0
+        mulps   xmm1,xmm7
+        movaps  .dn,xmm1
+        shufps  xmm3,xmm3,11111001b
+        subss   xmm3,.z1
+        mulss   xmm3,xmm7
+        movss   .dz,xmm3
+        subps   xmm6,xmm5
+        mulps   xmm6,xmm7
+        mov      ebx,.lx1
+        cmp      bx,.x_min     ; clipping on function4
+        jge      @f
+        movzx    eax,word .x_min
+        sub      eax,ebx
+        cvtsi2ss xmm7,eax
+        shufps   xmm7,xmm7,0
+        mulss    xmm3,xmm7
+        mulps    xmm1,xmm7
+        mulps    xmm6,xmm7
+        addss    xmm3,.z1
+        addps    xmm1,.n1
+        addps    xmm6,xmm5
+        movsx    eax,word .x_min
+        movss    .z1,xmm3
+        movaps   .n1,xmm1
+        mov      dword .lx1,eax
+      @@:
+        movzx   eax,word .x_max
+        cmp     .lx2,eax
+        jl      @f
+        mov     .lx2,eax
+      @@:
+        movzx   eax,word[xres_var]
+        mul     dword .y
+        add     eax,.lx1
+        mov     .zbuff,esi
+        mov     .screen,edi
+        shl     eax,2
+        add     edi,eax
+        add     esi,eax
+        mov     ecx,.lx2
+        sub     ecx,.lx1
+        movd       xmm0,[vect_x]
+        punpcklwd  xmm0,[the_zero]
+        cvtdq2ps   xmm0,xmm0
+        movaps     .vect_t,xmm0
+   .ddraw:
+        xorps    xmm0,xmm0
+        movss    xmm2,.z1
+        movss    xmm5,.z1
+        movaps   .final_col,xmm0
+        addss    xmm2,[f1]
+        subss    xmm5,[f1]
+        cmpnltss xmm2,dword[esi]
+        cmpnltss xmm5,dword[esi]
+        pxor     xmm2,xmm5
+        movd     eax,xmm2
+        or       eax,eax
+        jz       .skips
+        movaps   xmm7,.n1
+        andps    xmm7,[zero_hgst_dd]
+        mulps    xmm7,xmm7 ; normalize
+        haddps   xmm7,xmm7
+        haddps   xmm7,xmm7
+        rsqrtps  xmm7,xmm7
+        mulps    xmm7,.n1
+        movaps   .cnv,xmm7
+        mov      ebx,point_light_coords
+        mov      edx,lights_aligned
+        xor      eax,eax
+  .nx_light:
+   pushad
+   cvtsi2ss xmm0,.lx1
+   cvtsi2ss xmm1,.y
+   movss    xmm2,.z1
+   movlhps  xmm0,xmm1
+   shufps   xmm0,xmm2,11001000b
+   subps    xmm0,[ebx] ; xmm0 - ray end, -> current vertex
+   movaps   xmm3,[ebx]
+   andps    xmm0,[zero_hgst_dd]
+   movaps   xmm1,xmm0
+   mulps    xmm0,xmm0
+   haddps   xmm0,xmm0
+   haddps   xmm0,xmm0
+   sqrtps   xmm0,xmm0
+   movss    .Rlen,xmm0
+   rcpps    xmm0,xmm0
+   mulps    xmm0,xmm1    ; xmm0 - normalized ray vector
+   andps    xmm0,[zero_hgst_dd]
+   movaps   .nray,xmm0
+   movaps   .r1,xmm3   ; ray orgin
+ if 0
+   movaps   xmm1,xmm3
+   call     calc_bounding_box
+   mov      .aabb_mask,eax
+end if
+   mov      edi,[triangles_ptr]
+   xor      ecx,ecx
+ .nx_tri:  ; next triangle
+   cmp     ecx,.cur_tri ; prevent self shadowing
+   je      .skipp
+ if 0
+   mov     edi,ecx
+   imul    edi,[i12]
+   add     edi,[triangles_ptr]
+   mov     eax,[edi]
+   mov     ebx,[edi+4]
+   mov     edx,[edi+8]
+   imul    eax,[i12]
+   imul    ebx,[i12]
+   imul    edx,[i12]
+   add     eax,[points_ptr]
+   add     ebx,[points_ptr]
+   add     edx,[points_ptr]
+   movups   xmm2,[eax]
+   movups   xmm3,[ebx]
+   movups   xmm4,[edx]
+   andps    xmm2,[sign_mask]
+   andps    xmm3,[sign_mask]
+   andps    xmm4,[sign_mask]
+   movmskps ebx,xmm4
+   cmpeqps  xmm2,xmm3
+   cmpeqps  xmm3,xmm4
+   andps    xmm2,xmm3
+   movmskps eax,xmm2
+   and      eax,111b
+   and      ebx,111b
+   cmp      eax,111b
+   jne      @f
+   bt       .aabb_mask,ebx
+   jnc      .skipp
+  @@:
+end if
+   mov     edi,ecx
+   imul    edi,[i12]
+   add     edi,[triangles_ptr]
+   mov     eax,[edi]
+   mov     ebx,[edi+4]
+   mov     edx,[edi+8]
+   imul    eax,[i12]
+   imul    ebx,[i12]
+   imul    edx,[i12]
+   add     eax,[points_rotated_ptr]
+   add     ebx,[points_rotated_ptr]
+   add     edx,[points_rotated_ptr]
+   movups   xmm2,[eax]
+   movups   xmm3,[ebx]
+   movups   xmm4,[edx]
+   addps    xmm2,.vect_t
+   addps    xmm3,.vect_t
+   addps    xmm4,.vect_t
+;intersect_tri: procs header
+; in:
+;     xmm0 - ray direction  ; should be normalized
+;     xmm1 - ray orgin
+;     xmm2 - tri vert1
+;     xmm3 - tri vert2
+;     xmm4 - tri vert3
+;     if  eax = 1 - intersction with edge
+;        xmm6 - edge lenght
+;     if  eax = 0 - intersect with ray (classic)
+; out:
+;     eax  = 1 - intersection occured
+;     xmm0 - float lo -> hi = t, v, u, ...
+   movss     xmm6,.Rlen
+   movaps    xmm0,.nray
+   movaps    xmm1,.r1
+   subss     xmm6,[the_one]
+   mov       eax,1
+   push      ecx
+   call      intersect_tri
+   pop       ecx
+   cmp       eax,1
+   je        .inter
+ .skipp:
+ .skp:
+   inc       ecx
+   cmp       ecx,[triangles_count_var]
+   jnz       .nx_tri
+;   jz        .do_process
+;   comiss  xmm0,.Rlen
+;   jl      .inter
+   popad
+ .do_process:
+        movaps    xmm5,.nray  ;[edx]
+        andps     xmm5,[zero_hgst_dd]  ; global
+        mulps    xmm5,.cnv  ;.lv   ; last dword should be zeroed
+ ;       andps     xmm5,[sign_z]   ; global
+        haddps   xmm5,xmm5
+        haddps   xmm5,xmm5
+        andps     xmm5,[abs_mask]  ; global
+        movaps   xmm7,xmm5
+        mulps    xmm7,xmm7
+        mulps    xmm7,xmm7
+        mulps    xmm5,[edx+16]
+        mulps    xmm7,xmm7
+        mulps    xmm7,xmm7
+        mulps    xmm7,[edx+48]
+        addps    xmm5,xmm7
+        minps    xmm5,[mask_255f]  ; global
+        maxps    xmm5,.final_col     ;  addps  maxps
+        movaps   .final_col,xmm5
+        jmp     .nx_loop
+  .inter:
+        popad
+      .nx_loop:
+      ;  add      edx,64    ; unncomment to achive 3 lights
+      ;  add      ebx,16
+      ;  cmp      edx,lights_aligned_end    ; global
+      ;  jnz      .nx_light
+        movaps    xmm1,.final_col
+        cvtps2dq  xmm1,xmm1
+        packssdw  xmm1,xmm1
+        packuswb  xmm1,xmm1
+        movd      [edi],xmm1
+  .skips:
+        movaps   xmm0,.n1
+        movss    xmm2,.z1
+        add      edi,4
+        add      esi,4
+        add      dword .lx1,1
+        addps    xmm0,.dn
+        addss    xmm2,.dz
+        movaps   .n1,xmm0
+        movss    .z1,xmm2
+        dec      ecx
+        jnz      .ddraw
+  .end_rp_line:
+        add      esp,320
+        pop      ebp
+ret

 /programs/demos/view3ds/a_procs.inc
 ,3 → 1,200
+if Ext > SSE2
+ ;--------------------------------------------------------------------
+init_point_lights:
+  ;      mov       eax,1000
+  ;      cvtsi2ss  xmm1,eax
+  ;      shufps    xmm1,xmm1,11000000b
+  ;      mov       esi,lights_aligned
+  ;      mov       edi,point_light_coords
+  ;      mov       ecx,3
+  ;    @@:
+  ;      movaps    xmm0,[esi]
+  ;      addps     xmm0,[f05xz]
+  ;      mulps     xmm0,xmm1
+  ;      movaps    [edi],xmm0
+  ;      add       esi,64
+   ;     add       edi,16
+   ;     loop      @b
+      mov       ecx,3
+      mov       edi,point_light_coords
+    @@:
+      push      ecx
+      xor       ecx,ecx
+      movzx     edx,word[size_x_var]
+      call      random
+      cvtsi2ss  xmm0,eax
+      movss     [edi],xmm0
+      xor       ecx,ecx
+      movzx     edx,word[size_x_var]
+      call      random
+      cvtsi2ss  xmm0,eax
+      movss     [edi+4],xmm0
+   ;   movzx     ebx,word[size_x_var]
+   ;   shl       ebx,2
+   ;   neg       ebx
+      mov       ecx,-1900
+   ;   sub       ecx,100
+      mov       edx,-600
+      call      random
+      cvtsi2ss  xmm0,eax
+      movss     [edi+8],xmm0
+   ;   mov       dword[edi+8],-1700.0
+      mov       [edi+12],dword 0
+      add       edi,16
+      pop       ecx
+      loop      @b
+ret
+;------------------------------------------------------------------
+intersect_tri: ; Moeller-Trumbore method
+; in:
+;     xmm0 - ray direction  ; should be normalized
+;     xmm1 - ray orgin
+;     xmm2 - tri vert1
+;     xmm3 - tri vert2
+;     xmm4 - tri vert3
+;     if  eax = 1 - intersction with edge
+;        xmm6 - edge lenght
+;     if  eax = 0 - intersect with ray (classic)
+; out:
+;     eax  = 1 - intersection occured
+;     xmm0 - float lo -> hi = t, v, u, ...
+      push    ebp
+      mov     ebp,esp
+      and     ebp,-16
+      sub     esp,220
+       .dir    equ [ebp-16]
+       .origin equ [ebp-32]
+       .ta     equ [ebp-48]
+       .tb     equ [ebp-64]
+       .tc     equ [ebp-80]
+       .tvec   equ [ebp-96]
+       .pvec   equ [ebp-112]
+       .qvec   equ [ebp-128]
+       .e1     equ [ebp-128-16]
+       .ift    equ dword[ebp-152]
+       .invdet equ [ebp-156]
+       .det    equ [ebp-160]
+       .ed_l   equ [ebp-164]
+       .u      equ [ebp-168]
+       .v      equ [ebp-172]
+       .t      equ [ebp-176]
+       .e2     equ [ebp-192]
+      movaps   .dir,xmm0
+      movaps   .origin,xmm1
+      movaps   .ta,xmm2
+      movaps   .tb,xmm3
+      movaps   .tc,xmm4
+      mov      .ift,eax
+      movss    .ed_l,xmm6
+      subps    xmm3,xmm2
+      subps    xmm4,xmm2
+      andps    xmm3,[zero_hgst_dd]
+      andps    xmm4,[zero_hgst_dd]
+      movaps   .e1,xmm3
+      movaps   .e2,xmm4
+      lea      esi,.dir
+      lea      edi,.e2
+      lea      ebx,.pvec
+      call     cross_aligned
+      movaps   xmm0,.e1
+      mulps    xmm0,.pvec
+ ;     andps    xmm0,[zero_hgst_dd]
+      haddps   xmm0,xmm0
+      haddps   xmm0,xmm0
+      movss    .det,xmm0
+;      cmpnless xmm0,[eps]
+;      movd     eax,xmm0
+;      or       eax,eax
+;      jz       @f
+      comiss   xmm0,[eps]
+      jl       @f
+      rcpss    xmm0,.det
+      movss    .invdet,xmm0
+      movaps   xmm0,.origin
+      subps    xmm0,.ta
+      andps    xmm0,[zero_hgst_dd]
+      movaps   .tvec,xmm0
+      mulps    xmm0,.pvec
+      haddps   xmm0,xmm0
+      haddps   xmm0,xmm0
+      mulss    xmm0,.invdet
+      movss    xmm1,xmm0
+      movss    .u,xmm0
+      cmpnless xmm1,[epsone]
+      cmpnless xmm0,[epsminus]
+      pxor     xmm1,xmm0
+      movd     eax,xmm1
+      or       eax,eax
+      jz       @f
+      lea      esi,.tvec
+      lea      edi,.e1
+      lea      ebx,.qvec
+      call     cross_aligned
+      movaps   xmm0,.dir
+      mulps    xmm0,.qvec
+      haddps   xmm0,xmm0
+      haddps   xmm0,xmm0
+      mulss    xmm0,.invdet
+      movss    .v,xmm0
+      movss    xmm1,xmm0
+      addss    xmm1,.u
+      cmpnless xmm1,[epsone]
+      cmpnless xmm0,[epsminus]
+      pxor     xmm1,xmm0
+      movd     eax,xmm1
+      or       eax,eax
+      jz       @f
+      movaps   xmm1,.e2
+      mulps    xmm1,.qvec
+      haddps   xmm1,xmm1
+      haddps   xmm1,xmm1
+      mulss    xmm1,.invdet
+      movss    .t,xmm1
+   ;   cmpnless xmm1,[eps]
+   ;   movmskps eax,xmm1
+   ;   test     eax,1
+   ;   jz       @f
+      comiss   xmm1,[eps]
+      jl       @f
+      mov      eax,1
+      cmp      .ift,0
+      je       .end       ; ok intersect occured, no edge cause
+      movss    xmm0,.t      ; else check with edge lenght
+    ;  movss    xmm1,.t
+      cmpnless xmm0,[eps]
+      cmpnless xmm1,.ed_l
+      xorps    xmm0,xmm1
+      movd     ebx,xmm0
+      or       ebx,ebx
+      jz       @f
+ ;     mov      eax,1
+ ;     movaps   xmm0,.t
+      jmp      .end
+   @@:
+      xor      eax,eax
+    .end:
+      movaps   xmm0,.t
+      add      esp,220
+      pop      ebp
+ret
+end if
 ;===============================================================
 do_edges_list:
     push    ebp
 ,13 → 420,18
 do_sinus:
+;in - ax - render mode
    .x      equ  [ebp-8]
    .y      equ  [ebp-12]
    .new_y  equ  [ebp-16]
    .temp   equ  [ebp-20]
+   .dr_f   equ  word[ebp-22]
    push    ebp
    mov     ebp,esp
-   sub     esp,64
+   sub     esp,30
+   mov     .dr_f,ax
    mov     dword .x,0
    mov     dword .y,0
    mov     esi,[screen_ptr]
 ,53 → 445,20
    cld
    rep     stosd
    pop     edi
-;   movzx   eax,[sinus_flag]
-;   mov     edx,10
-;   mul     edx
-;   mov     [sin_amplitude],eax
-;   mov     [sin_frq],eax
    fninit
-;if Ext = SSE2
-;   movups  xmm1,[const0123]   ; xmm1 - init values
-;   mov     eax,0x000000ff
-;   movd    xmm2,eax
-;   shufps  xmm2,xmm2,0        ; xmm2 - mask value
-;   mov     eax,4
-;   movd    xmm3,eax
-;   shufps  xmm3,xmm3,0
  .again:
-if  0
    fild    dword .x
-   fidiv   [sin_frq]
-   fsin
-   fimul   [sin_amplitude]
-   fiadd   dword .y
-   fistp   dword .new_y
-else
-   fild    dword .x
    fmul    [sin_frq]
    fistp   dword .temp
    mov     eax, .temp
-;   mov     bx, [angle_x]
-;   add     bx, [angle_y]
-;   movzx   ebx,bx
-;   shr     ebx,1       ; change phase
-;   add     eax,ebx
    and     eax, 0x000000ff
-;   cdq
- ;  mul     [sin_frq]
-;   and      eax,0x000000ff
-;   and     ax,0x00ff
-;   cwde
    fld     dword [sin_tab+eax*4]
    fimul   dword [sin_amplitude]
    fiadd   dword .y
    fistp   dword .new_y
-end if
    mov     eax,.new_y
    or      eax,eax
    jl      .skip
 ,12 → 467,11
    jg      .skip
    movzx   edx,word[size_x_var]
    mul     edx
-;   shl     eax,9
    add     eax,dword .x
    lea     ebx,[eax*3]
-   cmp     [dr_flag],12 ; 32 bit col cause
-   jl      @f
+   cmp     .dr_f,12 ; 32 bit col cause
+   jb      @f
    add     ebx,eax
   @@:
    mov     eax,[esi]
 ,8 → 478,8
    mov     [edi+ebx],eax
  .skip:
    add     esi,3
-   cmp     [dr_flag],12
-   jl      @f
+   cmp     .dr_f,12
+   jb      @f
    inc     esi
   @@:
    inc     dword .x
 ,8 → 498,8
    movzx   ecx,word[size_x_var]
    movzx   eax,word[size_y_var]
    imul    ecx,eax
-   cmp     [dr_flag],12
-   jge     @f
+   cmp     .dr_f,12
+   jae     @f
    lea     ecx,[ecx*3]
    shr     ecx,2
  ;  mov     ecx,SIZE_X*SIZE_Y*3/4
 ,7 → 545,19
 ret
 do_emboss:   ; sse2 version only
+; in ax - render model
+   push  ebp
+   mov   ebp,esp
+   sub   esp,4
+   .dr_mod  equ word[ebp-2]
+   mov    .dr_mod,ax
 if Ext >= SSE2
  movzx ecx,[bumps_deep_flag]
  inc   ecx
  call  blur_screen    ;blur n times
 ,13 → 572,13
  sub   ecx,ebx
  mov   esi,[screen_ptr]
  mov   edi,[Zbuffer_ptr]
- cmp   [dr_flag],12
+ cmp   .dr_mod,11
  jge   @f
  lea   ebx,[ebx*3]
- jmp   .f
+ jmp   .gf
 @@:
  shl   ebx,2
-.f:
+.gf:
  mov   edx,esi
  add   esi,ebx
  lea   ebx,[ebx+esi]
 ,7 → 585,7
  pxor  xmm0,xmm0
  push  eax
 .emb:
- cmp   [dr_flag],12
+ cmp    .dr_mod ,11
  jge   @f
  movlps     xmm1,[esi+3]
  movhps     xmm1,[esi+6]
 ,14 → 622,7
  pmaxsw      xmm1,xmm7
  pmaxsw      xmm1,xmm6
-if 0
- movaps      xmm7,xmm3
- movaps      xmm6,xmm3
- psrlq       xmm7,2*8
- psrlq       xmm6,4*8
- pmaxsw      xmm3,xmm7
- pmaxsw      xmm3,xmm6
-end if
  pmaxsw      xmm1,xmm3
  movd        eax,xmm1
 ,7 → 642,7
  mov        eax,[eax]
  mov        [edi+4],eax
- cmp    [dr_flag],12
+ cmp    .dr_mod,11
  jl     @f
  add    esi,2
  add    ebx,2
 ,7 → 660,7
  pop    ecx  ;,eax
  mov    edi,[screen_ptr]
  mov    esi,[Zbuffer_ptr]
- cmp    [dr_flag],12
+ cmp    .dr_mod,11
  jge    .e
 @@:
  movsd
 ,6 → 671,11
 end if
+   mov   esp,ebp
+   pop   ebp
 ret
 ;align 16

 /programs/demos/view3ds/asc.inc
 ,815 → 1,815
-; Files *.asc routines by Maciej Guba
-; Thanks to Reverend for integer/float/ascii conversion examples
-read_asc:
-    mov  eax,[fptr]
- .find_vert:
-    cmp  dword[eax],'Vert'
-    je   @f
-    inc  eax
-    jmp  .find_vert
-  @@:
-    add   eax,4
-    cmp   dword[eax],'ices'
-    jne   .find_vert
-    add   eax,3
-  @@:
-    inc   eax
-    cmp   byte[eax],'0'    ; search end of ascii number of vertices string
-    jb    @b
-    cmp   byte[eax],'9'
-    ja    @b
-;    eax - start ascii number
-  @@:
-    inc    eax
-    cmp    byte[eax],'0'
-    jb    .convert1
-    cmp    byte[eax],'9'
-    ja    .convert1
-    jmp   @b
-  .convert1:
-    dec   eax
-    mov   ebx,eax
-    push  eax
-    call  ascii_to_integer
-    mov   [points_count_var],edx
-    pop   eax
-  @@:
-    inc   eax
-    cmp   dword[eax],'Face'
-    jne   @b
-    add   eax,3
-  @@:
-    inc   eax
-    cmp   byte[eax],'0'
-    jb    @b
-    cmp   byte[eax],'9'
-    ja    @b
-   ;    eax - start ascii number
-  @@:
-    inc    eax
-    cmp    byte[eax],'0'
-    jb    .convert2
-    cmp    byte[eax],'9'
-    ja    .convert2
-    jmp   @b
-   ;  eax - end ascii number
-  .convert2:
-    dec   eax
-    mov   ebx,eax
-    push  eax
-    call  ascii_to_integer
-    mov   [triangles_count_var],edx
-    pop   eax
-  @@:
-    inc   eax
-    cmp   dword[eax],'Vert'
-    jnz   @b
-    inc   eax
-    mov   edi,[points_ptr]
-    xor   ebx,ebx
-   .decode_vertices:
-    push   ebx
-  @@:
-    inc   eax
-    cmp   dword[eax],'Vert'
-    jne   @b
-    xor   ecx,ecx
-  .decode_coord:
-    push  ecx
-  @@:
-    inc   eax
-    mov   dl,byte[eax]
-    cmp   dl,byte[XYZpartices+ecx]
-    jne   @b
-  @@:
-    inc   eax
-    cmp   byte[eax],'.'
-    je    .readF
-    cmp   byte[eax],'-'
-    je    .readF
-    cmp   byte[eax],'0'
-    jb    @b
-    cmp   byte[eax],'9'
-    ja    @b
-.readF: ; read float
-    mov   esi,eax
-    push  eax
-    push  ecx
-    call   atof   ; st0 - desired dword float
-    pop   ecx
-    pop   eax
-    fstp  dword[edi]
-    add   edi,4
-    pop   ecx
-    inc   ecx
-    cmp   ecx,3
-    jne   .decode_coord
-    pop   ebx
-    inc   ebx
-    cmp   ebx,[points_count_var]
-    jne   .decode_vertices
-    mov   dword[edi],-1
-    mov   esi,eax
-   @@:
-    inc   esi
-    cmp   dword[esi],'Face'
-    jne   @b
-    xor   edx,edx
-    mov   edi,[triangles_ptr]
-    cld
- .decode_face:
-    push  edx
-   @@:
-    inc   esi
-    cmp   dword[esi],'Face'
-    jne   @b
- @@:
-    inc   esi
-    cmp   byte[esi],'0'    ; face number start
-    jb    @b
-    cmp   byte[esi],'9'
-    ja    @b
-  @@:
-    inc    esi
-    cmp    byte[esi],'0'
-    jb    @f
-    cmp    byte[esi],'9'   ; face number end
-    ja    @f
-    jmp   @b
-  @@:
-    xor ecx,ecx
- .next_vertex_number:
-    push  ecx
-  @@:
-    inc   esi
-    cmp   byte[esi],'0'
-    jb    @b
-    cmp   byte[esi],'9'
-    ja    @b
-    ;    eax - start ascii number
-  @@:
-    inc    esi
-    cmp    byte[esi],'0'
-    jb    @f
-    cmp    byte[esi],'9'
-    ja    @f
-    jmp   @b
-   ;  eax - end ascii number
-  @@:
-    dec   esi
-    mov   ebx,esi
-    push  esi
-    call  ascii_to_integer
-    mov   eax,edx
-    stosd
-    pop   esi
-    add   esi,4
-    pop   ecx
-    inc   ecx
-    cmp   ecx,3
-    jne   .next_vertex_number
-    pop   edx
-    inc   edx
-    cmp   edx,[triangles_count_var]
-    jne   .decode_face
-    mov   dword[edi],-1 ;dword[triangles+ebx+2],-1   ; end mark
-    mov   eax,1  ;-> mark if ok
-ret
-ascii_to_integer:
-; in ---  [ebx] -> end of ascii string
-; out --  edx  -> desired number
-    xor    edx,edx
-    xor    ecx,ecx
-  .again:
-    movzx  eax,byte[ebx]
-    sub    al,'0'
-    cwde
-    push   edx
-    mul    dword[convert_muler+ecx]
-    pop    edx
-    add    edx,eax
-    dec    ebx
-    cmp    byte[ebx],'0'
-    jb     .end
-    cmp    byte[ebx],'9'
-    ja     .end
-    add    ecx,4
-    jmp    .again
-  @@:
-  .end:
-ret
-;===============================================================================
-;       ASCII to float conversion procedure
-;
-;       input:
-;       esi - pointer to string
-;
-;       output:
-;       st0 - number changed into float
-;
-;===============================================================================
-atof:
-        .string  equ ebp-4
-        push    ebp
-        mov     ebp,esp
-        sub     esp,32
-        push    eax ecx esi
-        mov     [.string],esi
-        fninit
-        fldz
-        fldz
-        cld
-        cmp     byte [esi], '-'
-        jnz     @F
-        inc     esi
-    @@:
-        xor     eax, eax
-        align   4
-  .loop.integer_part:
-        lodsb
-        cmp     al, '.'
-        jz      .mantisa
-        cmp     al,'0'
-        jb      .exit
-        cmp     al,'9'
-        ja      .exit
-        fimul   [i10]
-        sub     al, '0'
-        push    eax
-        fiadd   dword [esp]
-        add     esp, 4
-        jmp     .loop.integer_part
-  .mantisa:
-        xor      ecx, ecx
-        xor      eax, eax
-        cld
-        fxch     st1
-     @@:
-        lodsb
-        cmp     al,'0'
-        jb      .exit
-        cmp     al,'9'
-        ja      .exit
-        cmp     ecx,7*4
-        je      .exit           ; max 7 digits in mantisa
-        sub     al,'0'
-        push    eax
-        fild    dword[esp]
-        fidiv   dword[convert_muler+4+ecx]
-        faddp
-        add     esp,4
-        add     ecx,4
-        jmp     @b
-  .exit:
-        faddp
-        mov     eax, [.string]
-        cmp     byte [eax], '-'
-        jnz     @F
-        fchs
-    @@:
-        cld
-        stc                             ; always returns no error
-        pop     esi ecx eax
-        mov     esp,ebp
-        pop     ebp
-        ret
-itoa:              ; unsigned dword integer to ascii procedure
-;     in   eax - variable
-;          esi - Pointer to ascii string
-;   out    esi - desired ascii string
-;          edi - end of ascii string  - ptr to memory
- .temp_string  equ dword[ebp-36]
- .ptr          equ dword[ebp-40]
- .var          equ dword[ebp-44]
-           push  ecx
-           push  ebp
-           mov   ebp,esp
-           sub   esp,64
-           mov   .var,eax
-           mov   eax,-1
-           lea   edi,.temp_string
-           cld
-           mov   ecx,9
-           rep   stosd      ; make floor
-           mov   .ptr,esi
-           lea   edi,.temp_string
-           add   edi,34
-           std
-           xor   eax,eax
-           stosb               ; mark begin
-           mov   eax,.var
-           mov   esi,10
-        @@:
-           xor   edx,edx
-           div   esi
-           xchg  eax,edx
-           add   al,'0'
-           stosb
-           xchg  eax,edx
-           or    eax,eax
-           jnz   @b
-           stosb                     ; mark end
-           lea   esi,.temp_string
-           cld
-        @@:
-           lodsb
-           or   al,al
-           jnz  @b
-           mov  edi,.ptr
-         @@:
-           lodsb
-           stosb
-           or   al,al
-           jnz  @b
-           mov  esp,ebp
-           pop  ebp
-           pop  ecx
-ret
-if 1
-ftoa_mac:
-;   in  :  esi - pointer to  dword float
-;          edi - pointer to ascii string
-  .ptr_f     equ dword[ebp-4]
-  .sign      equ dword[ebp-8]   ;  0 ->   less than zero, 1 - otherwise
-  .ptr_ascii equ dword[ebp-12]
-  .integer   equ dword[ebp-20]
-  .fraction  equ dword[ebp-28]
-  .status_orginal equ word[ebp-32]
-  .status_changed equ word[ebp-34]
-         push     ecx
-         push     ebp
-         mov      ebp,esp
-         sub      esp,64
-        fninit
-        fnstcw  .status_orginal
-        mov     ax, .status_orginal
-        or      ax, 0000110000000000b
-        mov     .status_changed, ax
-        fldcw   .status_changed
-;       --------------------------------
-;       check if signed
-        xor     eax, eax
-        fld     dword[esi]
-        fst     .sign
-        test    .sign, 80000000h
-        setz    al
-        mov     .sign, eax
-         mov       .ptr_f,esi
-         mov       .ptr_ascii,edi
-         fabs
-         fld       st0
-         frndint
-         fist      .integer
-         fsubp     st1, st0
-         mov       eax,.integer
-         mov       esi,.ptr_ascii
-         call      itoa
-      ;  edi   ->   ptr to end of ascii string
-         dec       edi
-         mov       al,'.'
-         stosb
-        mov     ecx, 6   ; max 6 digits in fraction part
-    .loop:
-        fimul   [i10]
-        fld     st0
-        frndint
-        fist    .fraction
-        fsubp   st1, st0
-        mov     esi,edi
-        mov     eax,.fraction
-        add     al,'0'
-        stosb
-        ftst
-        fnstsw  ax
-        test    ax, 0100000000000000b
-        jz      @F
-        test    ax, 0000010100000000b
-        jz      .finish
-    @@:
-        loop    .loop
-if 0
-        fldcw   .status_orginal
-        fimul   [i10]
-        fist    .fraction
-    ;    mov     esi,edi
-        mov     eax,.fraction
-        add     al,'0'
-        stosb
-    ;    call    itoa
-;       --------------------------------
-;       restore previous values
-    .finish:
-      ;  fstp    st0
-         ffree     st
-         mov       eax,.fraction
-         mov       esi,edi
-;         call      itoa
-         add       al,'0'
-         stosb
-end if
- .finish:
-  ffree st
-         cmp       .sign,0
-         jnz       @f
-         mov       esi,.ptr_ascii
-         dec       esi
-         mov       byte[esi],'-'
-       @@:
-         mov       esp,ebp
-         pop       ebp
-         pop       ecx
-ret
-end if
-if 0
-;===============================================================================
-;       float to ASCII conversion procedure
-;
-;       input:
-;       buffer - pointer to memory where output will be saved
-;       precision - number of digits after dot
-;
-;       output:
-;       no immediate output
-;
-;       notes:
-;       separate integer and mantisa part with dot '.'
-;       so GOOD   123.456
-;          WRONG  123,456
-;
-;       coded by Reverend // HTB + RAG
-;===============================================================================
-proc    ftoa                    buffer, precision
-locals
-  status_original             dw ?
-  status_changed              dw ?
-  integer                     dd ?
-  mantisa                     dd ?
-  signed                      dd ?
-endl
-        push    eax ecx;edi ecx
-;       --------------------------------
-;       set correct precision
-        mov     eax, [precision]
-        cmp     eax, 51
-        jb      @F
-        mov     eax, 51
-    @@:
-        mov     [precision], eax
-;       --------------------------------
-;       change control wortd of fpu to prevent rounding
-        fnstcw  [status_original]
-        mov     ax, [status_original]
-        or      ax, 0000110000000000b
-        mov     [status_changed], ax
-        fldcw   [status_changed]
-;       --------------------------------
-;       check if signed
-        xor     eax, eax
-        fst     [signed]
-        test    [signed], 80000000h
-        setnz   al
-        mov     [signed], eax
-;       --------------------------------
-;       cut integer and mantisa separately
-        fld     st0
-        fld     st0                     ; st0 = x, st1 = x
-        frndint
-        fist    [integer]               ; st0 = x, st1 = x
-        fabs
-        fsubp   st1, st0                ; st0 = mantisa(x)
-;       --------------------------------
-;       save integer part in buffer
-    ;    mov     edi, [buffer]
-        mov     esi,[buffer]
-      ;  push    [signed]
-      ;  push    edi
-      ;  push    10
-      ;  push    [integer]
-        mov     eax,[integer]
-        call    itoa
- ;       add     edi, eax
-        mov     al, '.'
-        stosb
-        mov     esi,edi
-;       --------------------------------
-;       save mantisa part in buffer
-        mov     ecx, [precision]
-        dec     ecx
-    .loop:
-        fimul   [i10]
-        fld     st0
-        frndint
-        fist    [mantisa]
-        fsubp   st1, st0
-;        push    0
-;        push    edi
-;        push    10
-;        push    [mantisa]
-        mov     esi,edi
-        mov     eax,[mantisa]
-        call    itoa
-      ;  add     edi, eax
-        ftst
-        fnstsw  ax
-        test    ax, 0100000000000000b
-        jz      @F
-        test    ax, 0000010100000000b
-        jz      .finish
-    @@:
-        loop    .loop
-        fldcw   [status_original]
-        fimul   [i10]
-        fist    [mantisa]
-;        push    0
-;        push    edi
-;        push    10
-;        push    [mantisa]
-        mov     esi,edi
-        mov     eax,[mantisa]
-        call    itoa
-;       --------------------------------
-;       restore previous values
-    .finish:
-        fstp    st0
-        cmp     [signed],1
-        jnz   @f
-        mov   byte[buffer],'-'
-      @@:
-        stc
-        pop     ecx eax   ;edi eax
-        ret
-endp
-end if
-if 0
-write_asc:
-    .counter equ dword[ebp-4]
-        push  ebp
-        mov   ebp,esp
-        sub   esp,64
-        fninit
-        mov   edi,asc_file_buffer
-        mov   esi,asc_main_header
-        cld
-      @@:
-        lodsb
-        cmp   al,1
-        jz    @f
-        stosb
-        jmp   @b
-      @@:
-        mov   esi,asc_info_header
-      @@:
-        lodsb
-        cmp   al,1
-        jz    @f
-        stosb
-        jmp   @b
-     @@:
-        push   esi  ; -> position in header info
-        movzx  eax,[points_count_var]
-        mov    esi,edi
-        call   itoa              ; unsigned dword integer to ascii procedure
-        pop   esi
-        inc   esi
-      @@:
-        lodsb
-        cmp   al,1
-        jz    @f
-        stosb
-        jmp   @b
-      @@:
-        push  esi
-        movzx eax,[triangles_count_var]
-        mov   esi,edi
-        call  itoa
-        pop   esi
-        inc   esi
-      @@:
-        lodsb
-        cmp   al,1
-        jz    @f
-        stosb
-        jmp   @b
-      @@:
-;=============================================================
-;================vertex list parser===========================
-;=============================================================
-       xor   ecx,ecx
-     .again_vertex:
-       push  ecx
-       mov   esi,asc_one_vertex_formula
-     @@:
-       lodsb
-       cmp   al,1
-       jz    @f
-       stosb
-       jmp   @b
-     @@:
-       mov   eax,ecx
- ;      push  ecx
-       push  esi
-       mov   esi,edi
-       call  itoa
-       pop   esi
-;       pop   ecx
-       inc   esi
-       xor   ebx,ebx
-    .next_vertex_coef:
-       push  ebx
-     @@:
-       lodsb
-       cmp   al,1
-       jz    @f
-       stosb
-       jmp   @b
-     @@:
-    ; int3
-       push  esi
-       lea   esi,[ecx*3]
-       shl   esi,2
-       add   esi,points_r
-       add   esi,ebx
-     ;  int3
-       call  ftoa_mac
-     ;  std
-     ;  fld     dword[esi]
-     ;  pushad
-     ;  stdcall ftoa, edi, 30
-     ;  popad
-     ;  add   edi,20
-       pop   esi
-       pop   ebx
-       add   ebx,4
-       cmp   ebx,12
-       jnz   .next_vertex_coef
-     @@:
-       lodsb
-       cmp   al,1
-       jz    @f
-       stosb
-       jmp   @b
-     @@:
-       pop   ecx
-       inc   ecx
-       cmp   cx,[points_count_var]
-       jnz   .again_vertex
-   ;   mov  edi,[temp_edi]
-       mov  esi,asc_face_list_header
-      @@:
-        lodsb
-        cmp   al,1      ; all face header
-        jz    @f
-        stosb
-        jmp   @b
-      @@:
-           ;=====================================
-           ; ==============face list parser======
-           ;=====================================
-       xor   ecx,ecx
-     .again_face:
-       push  ecx
-       mov   .counter,ecx
-       mov   esi,asc_one_face_formula
-     @@:
-       lodsb
-       cmp   al,1
-       jz    @f
-       stosb
-       jmp   @b
-     @@:
-       mov   eax,ecx
-       push  esi
-       mov   esi,edi
-       call  itoa
-       pop   esi
-       inc   esi
-     @@:
-       lodsb
-       cmp   al,1
-       jz    @f
-       stosb
-       jmp   @b
-     @@:
-       xor    ebx,ebx
-      .next_face_index:
-        push   ebx
-        mov    ecx,.counter
-        lea    ecx,[ecx*3]
-        add    ecx,ecx
-        movzx  eax,word[triangles+ecx+ebx]
-        push   esi
-        mov    esi,edi
-        call   itoa
-        pop    esi
-       @@:
-        lodsb
-        cmp   al,1
-        jz    @f
-        stosb
-        jmp   @b
-      @@:
-        pop    ebx
-        add    ebx,2
-        cmp    ebx,6
-        jnz    .next_face_index
-;        push   esi
-        mov   esi,asc_material
-     @@:
-        lodsb
-        cmp   al,1
-        jz    @f
-        stosb
-        jmp   @b
-     @@:
-;        pop   esi
-         pop    ecx
-        inc    ecx
-        cmp    cx,[triangles_count_var]
-        jnz    .again_face
-        ; write file
-    sub  edi,asc_file_buffer
-  ;  mov  [file_buffer+2],edi
-    mov  [FileSize],edi
-    invoke  CreateFile,asc_file_name, GENERIC_WRITE, 0, 0,CREATE_ALWAYS, FILE_ATTRIBUTE_NORMAL, 0
-    mov  [hfile],eax
-    invoke  WriteFile,eax,asc_file_buffer,[FileSize], byteswritten, 0
-    invoke  CloseHandle,[hfile]
-    mov  esp,ebp
-    pop  ebp
-ret
-end if
-if 0
-asc_file_buffer:
-              rd 65536
-ascii_string  rb 50
-ftoa_muler dd 1000000000
-file_size dd ?
-file_handle dd ?
-end if
-if 0
-convert_muler:
-dd 1, 10, 100, 1000, 10000
-XYZpartices:
-db 'X','Y','Z'
-i10 dw 10
-points_count_var dd ?
-triangles_count_var dd ?
-points rb 100
-triangles rb 100
-asc_file:
-file "2TORUS.ASC"
+; Files *.asc routines by Maciej Guba
+; Thanks to Reverend for integer/float/ascii conversion examples
+read_asc:
+    mov  eax,[fptr]
+ .find_vert:
+    cmp  dword[eax],'Vert'
+    je   @f
+    inc  eax
+    jmp  .find_vert
+  @@:
+    add   eax,4
+    cmp   dword[eax],'ices'
+    jne   .find_vert
+    add   eax,3
+  @@:
+    inc   eax
+    cmp   byte[eax],'0'    ; search end of ascii number of vertices string
+    jb    @b
+    cmp   byte[eax],'9'
+    ja    @b
+;    eax - start ascii number
+  @@:
+    inc    eax
+    cmp    byte[eax],'0'
+    jb    .convert1
+    cmp    byte[eax],'9'
+    ja    .convert1
+    jmp   @b
+  .convert1:
+    dec   eax
+    mov   ebx,eax
+    push  eax
+    call  ascii_to_integer
+    mov   [points_count_var],edx
+    pop   eax
+  @@:
+    inc   eax
+    cmp   dword[eax],'Face'
+    jne   @b
+    add   eax,3
+  @@:
+    inc   eax
+    cmp   byte[eax],'0'
+    jb    @b
+    cmp   byte[eax],'9'
+    ja    @b
+   ;    eax - start ascii number
+  @@:
+    inc    eax
+    cmp    byte[eax],'0'
+    jb    .convert2
+    cmp    byte[eax],'9'
+    ja    .convert2
+    jmp   @b
+   ;  eax - end ascii number
+  .convert2:
+    dec   eax
+    mov   ebx,eax
+    push  eax
+    call  ascii_to_integer
+    mov   [triangles_count_var],edx
+    pop   eax
+  @@:
+    inc   eax
+    cmp   dword[eax],'Vert'
+    jnz   @b
+    inc   eax
+    mov   edi,[points_ptr]
+    xor   ebx,ebx
+   .decode_vertices:
+    push   ebx
+  @@:
+    inc   eax
+    cmp   dword[eax],'Vert'
+    jne   @b
+    xor   ecx,ecx
+  .decode_coord:
+    push  ecx
+  @@:
+    inc   eax
+    mov   dl,byte[eax]
+    cmp   dl,byte[XYZpartices+ecx]
+    jne   @b
+  @@:
+    inc   eax
+    cmp   byte[eax],'.'
+    je    .readF
+    cmp   byte[eax],'-'
+    je    .readF
+    cmp   byte[eax],'0'
+    jb    @b
+    cmp   byte[eax],'9'
+    ja    @b
+.readF: ; read float
+    mov   esi,eax
+    push  eax
+    push  ecx
+    call   atof   ; st0 - desired dword float
+    pop   ecx
+    pop   eax
+    fstp  dword[edi]
+    add   edi,4
+    pop   ecx
+    inc   ecx
+    cmp   ecx,3
+    jne   .decode_coord
+    pop   ebx
+    inc   ebx
+    cmp   ebx,[points_count_var]
+    jne   .decode_vertices
+    mov   dword[edi],-1
+    mov   esi,eax
+   @@:
+    inc   esi
+    cmp   dword[esi],'Face'
+    jne   @b
+    xor   edx,edx
+    mov   edi,[triangles_ptr]
+    cld
+ .decode_face:
+    push  edx
+   @@:
+    inc   esi
+    cmp   dword[esi],'Face'
+    jne   @b
+ @@:
+    inc   esi
+    cmp   byte[esi],'0'    ; face number start
+    jb    @b
+    cmp   byte[esi],'9'
+    ja    @b
+  @@:
+    inc    esi
+    cmp    byte[esi],'0'
+    jb    @f
+    cmp    byte[esi],'9'   ; face number end
+    ja    @f
+    jmp   @b
+  @@:
+    xor ecx,ecx
+ .next_vertex_number:
+    push  ecx
+  @@:
+    inc   esi
+    cmp   byte[esi],'0'
+    jb    @b
+    cmp   byte[esi],'9'
+    ja    @b
+    ;    eax - start ascii number
+  @@:
+    inc    esi
+    cmp    byte[esi],'0'
+    jb    @f
+    cmp    byte[esi],'9'
+    ja    @f
+    jmp   @b
+   ;  eax - end ascii number
+  @@:
+    dec   esi
+    mov   ebx,esi
+    push  esi
+    call  ascii_to_integer
+    mov   eax,edx
+    stosd
+    pop   esi
+    add   esi,4
+    pop   ecx
+    inc   ecx
+    cmp   ecx,3
+    jne   .next_vertex_number
+    pop   edx
+    inc   edx
+    cmp   edx,[triangles_count_var]
+    jne   .decode_face
+    mov   dword[edi],-1 ;dword[triangles+ebx+2],-1   ; end mark
+    mov   eax,1  ;-> mark if ok
+ret
+ascii_to_integer:
+; in ---  [ebx] -> end of ascii string
+; out --  edx  -> desired number
+    xor    edx,edx
+    xor    ecx,ecx
+  .again:
+    movzx  eax,byte[ebx]
+    sub    al,'0'
+    cwde
+    push   edx
+    mul    dword[convert_muler+ecx]
+    pop    edx
+    add    edx,eax
+    dec    ebx
+    cmp    byte[ebx],'0'
+    jb     .end
+    cmp    byte[ebx],'9'
+    ja     .end
+    add    ecx,4
+    jmp    .again
+  @@:
+  .end:
+ret
+;===============================================================================
+;       ASCII to float conversion procedure
+;
+;       input:
+;       esi - pointer to string
+;
+;       output:
+;       st0 - number changed into float
+;
+;===============================================================================
+atof:
+        .string  equ ebp-4
+        push    ebp
+        mov     ebp,esp
+        sub     esp,32
+        push    eax ecx esi
+        mov     [.string],esi
+        fninit
+        fldz
+        fldz
+        cld
+        cmp     byte [esi], '-'
+        jnz     @F
+        inc     esi
+    @@:
+        xor     eax, eax
+        align   4
+  .loop.integer_part:
+        lodsb
+        cmp     al, '.'
+        jz      .mantisa
+        cmp     al,'0'
+        jb      .exit
+        cmp     al,'9'
+        ja      .exit
+        fimul   [i10]
+        sub     al, '0'
+        push    eax
+        fiadd   dword [esp]
+        add     esp, 4
+        jmp     .loop.integer_part
+  .mantisa:
+        xor      ecx, ecx
+        xor      eax, eax
+        cld
+        fxch     st1
+     @@:
+        lodsb
+        cmp     al,'0'
+        jb      .exit
+        cmp     al,'9'
+        ja      .exit
+        cmp     ecx,7*4
+        je      .exit           ; max 7 digits in mantisa
+        sub     al,'0'
+        push    eax
+        fild    dword[esp]
+        fidiv   dword[convert_muler+4+ecx]
+        faddp
+        add     esp,4
+        add     ecx,4
+        jmp     @b
+  .exit:
+        faddp
+        mov     eax, [.string]
+        cmp     byte [eax], '-'
+        jnz     @F
+        fchs
+    @@:
+        cld
+        stc                             ; always returns no error
+        pop     esi ecx eax
+        mov     esp,ebp
+        pop     ebp
+        ret
+itoa:              ; unsigned dword integer to ascii procedure
+;     in   eax - variable
+;          esi - Pointer to ascii string
+;   out    esi - desired ascii string
+;          edi - end of ascii string  - ptr to memory
+ .temp_string  equ dword[ebp-36]
+ .ptr          equ dword[ebp-40]
+ .var          equ dword[ebp-44]
+           push  ecx
+           push  ebp
+           mov   ebp,esp
+           sub   esp,64
+           mov   .var,eax
+           mov   eax,-1
+           lea   edi,.temp_string
+           cld
+           mov   ecx,9
+           rep   stosd      ; make floor
+           mov   .ptr,esi
+           lea   edi,.temp_string
+           add   edi,34
+           std
+           xor   eax,eax
+           stosb               ; mark begin
+           mov   eax,.var
+           mov   esi,10
+        @@:
+           xor   edx,edx
+           div   esi
+           xchg  eax,edx
+           add   al,'0'
+           stosb
+           xchg  eax,edx
+           or    eax,eax
+           jnz   @b
+           stosb                     ; mark end
+           lea   esi,.temp_string
+           cld
+        @@:
+           lodsb
+           or   al,al
+           jnz  @b
+           mov  edi,.ptr
+         @@:
+           lodsb
+           stosb
+           or   al,al
+           jnz  @b
+           mov  esp,ebp
+           pop  ebp
+           pop  ecx
+ret
+if 1
+ftoa_mac:
+;   in  :  esi - pointer to  dword float
+;          edi - pointer to ascii string
+  .ptr_f     equ dword[ebp-4]
+  .sign      equ dword[ebp-8]   ;  0 ->   less than zero, 1 - otherwise
+  .ptr_ascii equ dword[ebp-12]
+  .integer   equ dword[ebp-20]
+  .fraction  equ dword[ebp-28]
+  .status_orginal equ word[ebp-32]
+  .status_changed equ word[ebp-34]
+         push     ecx
+         push     ebp
+         mov      ebp,esp
+         sub      esp,64
+        fninit
+        fnstcw  .status_orginal
+        mov     ax, .status_orginal
+        or      ax, 0000110000000000b
+        mov     .status_changed, ax
+        fldcw   .status_changed
+;       --------------------------------
+;       check if signed
+        xor     eax, eax
+        fld     dword[esi]
+        fst     .sign
+        test    .sign, 80000000h
+        setz    al
+        mov     .sign, eax
+         mov       .ptr_f,esi
+         mov       .ptr_ascii,edi
+         fabs
+         fld       st0
+         frndint
+         fist      .integer
+         fsubp     st1, st0
+         mov       eax,.integer
+         mov       esi,.ptr_ascii
+         call      itoa
+      ;  edi   ->   ptr to end of ascii string
+         dec       edi
+         mov       al,'.'
+         stosb
+        mov     ecx, 6   ; max 6 digits in fraction part
+    .loop:
+        fimul   [i10]
+        fld     st0
+        frndint
+        fist    .fraction
+        fsubp   st1, st0
+        mov     esi,edi
+        mov     eax,.fraction
+        add     al,'0'
+        stosb
+        ftst
+        fnstsw  ax
+        test    ax, 0100000000000000b
+        jz      @F
+        test    ax, 0000010100000000b
+        jz      .finish
+    @@:
+        loop    .loop
+if 0
+        fldcw   .status_orginal
+        fimul   [i10]
+        fist    .fraction
+    ;    mov     esi,edi
+        mov     eax,.fraction
+        add     al,'0'
+        stosb
+    ;    call    itoa
+;       --------------------------------
+;       restore previous values
+    .finish:
+      ;  fstp    st0
+         ffree     st
+         mov       eax,.fraction
+         mov       esi,edi
+;         call      itoa
+         add       al,'0'
+         stosb
+end if
+ .finish:
+  ffree st
+         cmp       .sign,0
+         jnz       @f
+         mov       esi,.ptr_ascii
+         dec       esi
+         mov       byte[esi],'-'
+       @@:
+         mov       esp,ebp
+         pop       ebp
+         pop       ecx
+ret
+end if
+if 0
+;===============================================================================
+;       float to ASCII conversion procedure
+;
+;       input:
+;       buffer - pointer to memory where output will be saved
+;       precision - number of digits after dot
+;
+;       output:
+;       no immediate output
+;
+;       notes:
+;       separate integer and mantisa part with dot '.'
+;       so GOOD   123.456
+;          WRONG  123,456
+;
+;       coded by Reverend // HTB + RAG
+;===============================================================================
+proc    ftoa                    buffer, precision
+locals
+  status_original             dw ?
+  status_changed              dw ?
+  integer                     dd ?
+  mantisa                     dd ?
+  signed                      dd ?
+endl
+        push    eax ecx;edi ecx
+;       --------------------------------
+;       set correct precision
+        mov     eax, [precision]
+        cmp     eax, 51
+        jb      @F
+        mov     eax, 51
+    @@:
+        mov     [precision], eax
+;       --------------------------------
+;       change control wortd of fpu to prevent rounding
+        fnstcw  [status_original]
+        mov     ax, [status_original]
+        or      ax, 0000110000000000b
+        mov     [status_changed], ax
+        fldcw   [status_changed]
+;       --------------------------------
+;       check if signed
+        xor     eax, eax
+        fst     [signed]
+        test    [signed], 80000000h
+        setnz   al
+        mov     [signed], eax
+;       --------------------------------
+;       cut integer and mantisa separately
+        fld     st0
+        fld     st0                     ; st0 = x, st1 = x
+        frndint
+        fist    [integer]               ; st0 = x, st1 = x
+        fabs
+        fsubp   st1, st0                ; st0 = mantisa(x)
+;       --------------------------------
+;       save integer part in buffer
+    ;    mov     edi, [buffer]
+        mov     esi,[buffer]
+      ;  push    [signed]
+      ;  push    edi
+      ;  push    10
+      ;  push    [integer]
+        mov     eax,[integer]
+        call    itoa
+ ;       add     edi, eax
+        mov     al, '.'
+        stosb
+        mov     esi,edi
+;       --------------------------------
+;       save mantisa part in buffer
+        mov     ecx, [precision]
+        dec     ecx
+    .loop:
+        fimul   [i10]
+        fld     st0
+        frndint
+        fist    [mantisa]
+        fsubp   st1, st0
+;        push    0
+;        push    edi
+;        push    10
+;        push    [mantisa]
+        mov     esi,edi
+        mov     eax,[mantisa]
+        call    itoa
+      ;  add     edi, eax
+        ftst
+        fnstsw  ax
+        test    ax, 0100000000000000b
+        jz      @F
+        test    ax, 0000010100000000b
+        jz      .finish
+    @@:
+        loop    .loop
+        fldcw   [status_original]
+        fimul   [i10]
+        fist    [mantisa]
+;        push    0
+;        push    edi
+;        push    10
+;        push    [mantisa]
+        mov     esi,edi
+        mov     eax,[mantisa]
+        call    itoa
+;       --------------------------------
+;       restore previous values
+    .finish:
+        fstp    st0
+        cmp     [signed],1
+        jnz   @f
+        mov   byte[buffer],'-'
+      @@:
+        stc
+        pop     ecx eax   ;edi eax
+        ret
+endp
+end if
+if 0
+write_asc:
+    .counter equ dword[ebp-4]
+        push  ebp
+        mov   ebp,esp
+        sub   esp,64
+        fninit
+        mov   edi,asc_file_buffer
+        mov   esi,asc_main_header
+        cld
+      @@:
+        lodsb
+        cmp   al,1
+        jz    @f
+        stosb
+        jmp   @b
+      @@:
+        mov   esi,asc_info_header
+      @@:
+        lodsb
+        cmp   al,1
+        jz    @f
+        stosb
+        jmp   @b
+     @@:
+        push   esi  ; -> position in header info
+        movzx  eax,[points_count_var]
+        mov    esi,edi
+        call   itoa              ; unsigned dword integer to ascii procedure
+        pop   esi
+        inc   esi
+      @@:
+        lodsb
+        cmp   al,1
+        jz    @f
+        stosb
+        jmp   @b
+      @@:
+        push  esi
+        movzx eax,[triangles_count_var]
+        mov   esi,edi
+        call  itoa
+        pop   esi
+        inc   esi
+      @@:
+        lodsb
+        cmp   al,1
+        jz    @f
+        stosb
+        jmp   @b
+      @@:
+;=============================================================
+;================vertex list parser===========================
+;=============================================================
+       xor   ecx,ecx
+     .again_vertex:
+       push  ecx
+       mov   esi,asc_one_vertex_formula
+     @@:
+       lodsb
+       cmp   al,1
+       jz    @f
+       stosb
+       jmp   @b
+     @@:
+       mov   eax,ecx
+ ;      push  ecx
+       push  esi
+       mov   esi,edi
+       call  itoa
+       pop   esi
+;       pop   ecx
+       inc   esi
+       xor   ebx,ebx
+    .next_vertex_coef:
+       push  ebx
+     @@:
+       lodsb
+       cmp   al,1
+       jz    @f
+       stosb
+       jmp   @b
+     @@:
+    ; int3
+       push  esi
+       lea   esi,[ecx*3]
+       shl   esi,2
+       add   esi,points_r
+       add   esi,ebx
+     ;  int3
+       call  ftoa_mac
+     ;  std
+     ;  fld     dword[esi]
+     ;  pushad
+     ;  stdcall ftoa, edi, 30
+     ;  popad
+     ;  add   edi,20
+       pop   esi
+       pop   ebx
+       add   ebx,4
+       cmp   ebx,12
+       jnz   .next_vertex_coef
+     @@:
+       lodsb
+       cmp   al,1
+       jz    @f
+       stosb
+       jmp   @b
+     @@:
+       pop   ecx
+       inc   ecx
+       cmp   cx,[points_count_var]
+       jnz   .again_vertex
+   ;   mov  edi,[temp_edi]
+       mov  esi,asc_face_list_header
+      @@:
+        lodsb
+        cmp   al,1      ; all face header
+        jz    @f
+        stosb
+        jmp   @b
+      @@:
+           ;=====================================
+           ; ==============face list parser======
+           ;=====================================
+       xor   ecx,ecx
+     .again_face:
+       push  ecx
+       mov   .counter,ecx
+       mov   esi,asc_one_face_formula
+     @@:
+       lodsb
+       cmp   al,1
+       jz    @f
+       stosb
+       jmp   @b
+     @@:
+       mov   eax,ecx
+       push  esi
+       mov   esi,edi
+       call  itoa
+       pop   esi
+       inc   esi
+     @@:
+       lodsb
+       cmp   al,1
+       jz    @f
+       stosb
+       jmp   @b
+     @@:
+       xor    ebx,ebx
+      .next_face_index:
+        push   ebx
+        mov    ecx,.counter
+        lea    ecx,[ecx*3]
+        add    ecx,ecx
+        movzx  eax,word[triangles+ecx+ebx]
+        push   esi
+        mov    esi,edi
+        call   itoa
+        pop    esi
+       @@:
+        lodsb
+        cmp   al,1
+        jz    @f
+        stosb
+        jmp   @b
+      @@:
+        pop    ebx
+        add    ebx,2
+        cmp    ebx,6
+        jnz    .next_face_index
+;        push   esi
+        mov   esi,asc_material
+     @@:
+        lodsb
+        cmp   al,1
+        jz    @f
+        stosb
+        jmp   @b
+     @@:
+;        pop   esi
+         pop    ecx
+        inc    ecx
+        cmp    cx,[triangles_count_var]
+        jnz    .again_face
+        ; write file
+    sub  edi,asc_file_buffer
+  ;  mov  [file_buffer+2],edi
+    mov  [FileSize],edi
+    invoke  CreateFile,asc_file_name, GENERIC_WRITE, 0, 0,CREATE_ALWAYS, FILE_ATTRIBUTE_NORMAL, 0
+    mov  [hfile],eax
+    invoke  WriteFile,eax,asc_file_buffer,[FileSize], byteswritten, 0
+    invoke  CloseHandle,[hfile]
+    mov  esp,ebp
+    pop  ebp
+ret
+end if
+if 0
+asc_file_buffer:
+              rd 65536
+ascii_string  rb 50
+ftoa_muler dd 1000000000
+file_size dd ?
+file_handle dd ?
+end if
+if 0
+convert_muler:
+dd 1, 10, 100, 1000, 10000
+XYZpartices:
+db 'X','Y','Z'
+i10 dw 10
+points_count_var dd ?
+triangles_count_var dd ?
+points rb 100
+triangles rb 100
+asc_file:
+file "2TORUS.ASC"
 end if

 /programs/demos/view3ds/b_procs.inc
 ,22 → 737,33
 blur_screen:    ;blur n times  ; blur or fire
 ;in - ecx  times count
-;.counter equ dword[esp-4]
-.counter1 equ dword[esp-8]
+;     ax - render mode
+ .val      equ  dword[ebp-4]
+ .dr_model equ  word[ebp-6]
+ .fire     equ  dword[ebp-10]
 if Ext>=SSE2
          push       ebp
          mov        ebp,esp
-         push       dword 0x01010101
-         movss      xmm5,[esp]
+         sub        esp,10
+      ;   xorps      xmm5,xmm5
+      ;   or         edx,edx
+      ;   jz         @f
+         mov        .val,0x01010101
+         movss      xmm5,.val
          shufps     xmm5,xmm5,0
+      @@:
+         mov        .dr_model,ax
      .again_blur:
          push       ecx
          mov        edi,[screen_ptr]
          movzx      ecx,word[size_x_var]  ;SIZE_X*3/4
-         cmp         [dr_flag],12
+         cmp         .dr_model,11
          jge        @f
-         lea        ecx,[ecx*3+1]
+         lea        ecx,[ecx*3+3]
          shr        ecx,2
        @@:
 ,11 → 774,11
          movzx      ecx,word[size_y_var]
          sub        ecx,3
          imul       ecx,ebx
-         cmp        [dr_flag],12  ; 32 bit per pix cause
+         cmp        .dr_model,11  ; 32 bit per pix cause
          jge        @f
          lea        ecx,[ecx*3]
          shr        ecx,4
-         lea        ebx,[ebx *3]
+         lea        ebx,[ebx*3]
          jmp        .blr
        @@:
 ,7 → 792,7
          mov        ecx,edi
          sub        ecx,ebx
          movups     xmm1,[ecx]
-         cmp         [dr_flag],12
+         cmp        .dr_model,12
          jge        @f
          movups     xmm2,[edi-3]
          movups     xmm3,[edi+3]
 ,9 → 813,9
     end if
          xor        eax,eax
          movzx      ecx,word[size_x_var]
-         cmp        [dr_flag],12
+         cmp        .dr_model,11
          jge        @f
-         lea        ecx,[ecx*3]
+         lea        ecx,[ecx*3+3]
          shr        ecx,2
         @@:
     ;     mov        ecx,SIZE_X*3/4

 /programs/demos/view3ds/bump_cat.inc
 ,1132 → 1,1132
-;SIZE_X equ 350
-;SIZE_Y equ 350
-ROUND equ 8
-;TEX_X equ 512
-;TEX_Y equ 512
-;TEXTURE_SIZE EQU (512*512)-1
-;TEX_SHIFT EQU 9
-CATMULL_SHIFT equ 8
-;TEXTURE_SIZE EQU (TEX_X * TEX_Y)-1
-;Ext = NON
-;MMX = 1
-;NON = 0
-;------- Big thanks to Majuma (www.majuma.xt.pl) for absolutely great---
-;------- DOS 13h mode demos --------------------------------------------
-;------- Procedure draws bump triangle using Catmull Z-buffer algorithm-
-;------- (Z coordinate interpolation)-----------------------------------
-bump_triangle_z:
-;------------------in - eax - x1 shl 16 + y1 -----------
-;---------------------- ebx - x2 shl 16 + y2 -----------
-;---------------------- ecx - x3 shl 16 + y3 -----------
-;---------------------- edx - pointer to bump map ------
-;---------------------- esi - pointer to environment map
-;---------------------- edi - pointer to screen buffer--
-;---------------------- stack : bump coordinates--------
-;----------------------         environment coordinates-
-;----------------------         Z position coordinates--
-;----------------------         pointer io Z buffer-----
-;-- Z-buffer - filled with coordinates as dword --------
-;-- (Z coor. as word) shl CATMULL_SHIFT ----------------
-.b_x1   equ ebp+4   ; procedure don't save registers !!!
-.b_y1   equ ebp+6   ; each coordinate as word
-.b_x2   equ ebp+8
-.b_y2   equ ebp+10
-.b_x3   equ ebp+12
-.b_y3   equ ebp+14
-.e_x1   equ ebp+16
-.e_y1   equ ebp+18
-.e_x2   equ ebp+20
-.e_y2   equ ebp+22
-.e_x3   equ ebp+24
-.e_y3   equ ebp+26
-.z1     equ word[ebp+28]
-.z2     equ word[ebp+30]
-.z3     equ word[ebp+32]
-.z_buff equ dword[ebp+34]       ; pointer to Z-buffer
-.t_bmap equ dword[ebp-4]        ; pointer to bump map
-.t_emap equ dword[ebp-8]        ; pointer to e. map
-.x1     equ word[ebp-10]
-.y1     equ word[ebp-12]
-.x2     equ word[ebp-14]
-.y2     equ word[ebp-16]
-.x3     equ word[ebp-18]
-.y3     equ word[ebp-20]
-.dx12  equ dword[ebp-24]
-.dz12  equ      [ebp-28]
-.dbx12 equ dword[ebp-32]
-.dby12 equ      [ebp-36]
-.dex12 equ dword[ebp-40]
-.dey12 equ      [ebp-44]
-.dx13  equ dword[ebp-48]
-.dz13  equ      [ebp-52]
-.dbx13 equ dword[ebp-56]
-.dby13 equ      [ebp-60]
-.dex13 equ dword[ebp-64]
-.dey13 equ      [ebp-68]
-.dx23  equ dword[ebp-72]
-.dz23  equ      [ebp-76]
-.dbx23 equ dword[ebp-80]
-.dby23 equ      [ebp-84]
-.dex23 equ dword[ebp-88]
-.dey23 equ      [ebp-92]
-.cx1   equ dword[ebp-96]   ; current variables
-.cz1   equ      [ebp-100]
-.cx2   equ dword[ebp-104]
-.cz2   equ      [ebp-108]
-.cbx1  equ dword[ebp-112]
-.cby1  equ      [ebp-116]
-.cex1  equ dword[ebp-120]
-.cey1  equ      [ebp-124]
-.cbx2  equ dword[ebp-128]
-.cby2  equ      [ebp-132]
-.cex2  equ dword[ebp-136]
-.cey2  equ      [ebp-140]
-       mov     ebp,esp
-       push    edx        ; store bump map
-       push    esi        ; store e. map
-     ; sub     esp,120
- .sort3:                  ; sort triangle coordinates...
-       cmp     ax,bx
-       jle     .sort1
-       xchg    eax,ebx
-       mov     edx,dword[.b_x1]
-       xchg    edx,dword[.b_x2]
-       mov     dword[.b_x1],edx
-       mov     edx,dword[.e_x1]
-       xchg    edx,dword[.e_x2]
-       mov     dword[.e_x1],edx
-       mov     dx,.z1
-       xchg    dx,.z2
-       mov     .z1,dx
- .sort1:
-       cmp      bx,cx
-       jle      .sort2
-       xchg     ebx,ecx
-       mov      edx,dword[.b_x2]
-       xchg     edx,dword[.b_x3]
-       mov      dword[.b_x2],edx
-       mov      edx,dword[.e_x2]
-       xchg     edx,dword[.e_x3]
-       mov      dword[.e_x2],edx
-       mov     dx,.z2
-       xchg    dx,.z3
-       mov     .z2,dx
-       jmp      .sort3
- .sort2:
-       push     eax     ; store triangle coords in variables
-       push     ebx
-       push     ecx
-         mov      edx,80008000h  ; eax,ebx,ecx are ANDd together into edx which means that
-         and      edx,ebx        ; if *all* of them are negative a sign flag is raised
-         and      edx,ecx
-         and      edx,eax
-         test     edx,80008000h  ; Check both X&Y at once
-         jne      .loop23_done
-    ;   mov     edx,eax         ; eax,ebx,ecx are ORd together into edx which means that
-    ;   or      edx,ebx         ; if any *one* of them is negative a sign flag is raised
-    ;   or      edx,ecx
-    ;   test    edx,80000000h   ; Check only X
-    ;   jne     .loop23_done
-    ;   cmp     .x1,SIZE_X    ; {
-    ;   jg      .loop23_done
-    ;   cmp     .x2,SIZE_X     ; This can be optimized with effort
-    ;   jg      .loop23_done
-    ;   cmp     .x3,SIZE_X
-    ;   jg      .loop23_done    ; {
-       mov      bx,.y2       ; calc delta 12
-       sub      bx,.y1
-       jnz      .bt_dx12_make
-       mov      ecx,6
-       xor      edx,edx
-     @@:
-       push     edx   ;dword 0
-       loop     @b
-       jmp      .bt_dx12_done
- .bt_dx12_make:
-       mov      ax,.x2
-       sub      ax,.x1
-       cwde
-       movsx    ebx,bx
-       shl      eax,ROUND
-       cdq
-       idiv     ebx
- ;      mov      .dx12,eax
-       push      eax
-       mov     ax,.z2
-       sub     ax,.z1
-       cwde
-       shl     eax,CATMULL_SHIFT
-       cdq
-       idiv    ebx
-       push    eax
-if Ext>=SSE
-       sub       esp,16
-       cvtsi2ss  xmm3,ebx            ;rcps
-   ;    mov       eax,255
-       cvtsi2ss  xmm4,[i255d]
-       divss     xmm3,xmm4
-       rcpss     xmm3,xmm3
-    ;   mulss     xmm3,xmm4
-       shufps    xmm3,xmm3,0
-       movd      mm0,[.b_x1]
-       movd      mm1,[.b_x2]
-       movd      mm2,[.e_x1]
-       movd      mm3,[.e_x2]
-       pxor      mm4,mm4
-       punpcklwd mm0,mm4
-       punpcklwd mm1,mm4
-       punpcklwd mm2,mm4
-       punpcklwd mm3,mm4
-       psubd      mm1,mm0
-       psubd      mm3,mm2
- ;       cvtpi2ps  xmm0,mm0
- ;       movlhps   xmm0,xmm0
- ;       cvtpi2ps  xmm0,mm2
-       cvtpi2ps  xmm1,mm1
-       movlhps   xmm1,xmm1
-       cvtpi2ps  xmm1,mm3
- ;       subps     xmm1,xmm0
-;       cvtpi2ps  xmm0,mm3
-  ;     divps     xmm1,xmm3
-       mulps     xmm1,xmm3
-       shufps    xmm1,xmm1,10110001b
-       cvtps2pi  mm0,xmm1          ; mm0 -> 2 delta dwords
-       movhlps   xmm1,xmm1
-       cvtps2pi  mm1,xmm1
-       movq      .dey12,mm0
-       movq      .dby12,mm1
-else
-       mov      ax,word[.b_x2]
-       sub      ax,word[.b_x1]
-       cwde
-       shl      eax,ROUND
-       cdq
-       idiv     ebx
- ;      mov      .dbx12,eax
-       push      eax
-       mov      ax,word[.b_y2]
-       sub      ax,word[.b_y1]
-       cwde
-       shl      eax,ROUND
-       cdq
-       idiv     ebx
- ;      mov      .dby12,eax
-       push      eax
-       mov      ax,word[.e_x2]
-       sub      ax,word[.e_x1]
-       cwde
-       shl      eax,ROUND
-       cdq
-       idiv     ebx
- ;      mov      .dex12,eax
-       push      eax
-       mov      ax,word[.e_y2]
-       sub      ax,word[.e_y1]
-       cwde
-       shl      eax,ROUND
-       cdq
-       idiv     ebx
- ;      mov      .dey12,eax
-       push      eax
-end if
-   .bt_dx12_done:
-       mov      bx,.y3       ; calc delta13
-       sub      bx,.y1
-       jnz      .bt_dx13_make
-       mov      ecx,6
-       xor      edx,edx
-     @@:
-       push     edx   ;dword 0
-       loop     @b
-       jmp      .bt_dx13_done
- .bt_dx13_make:
-       mov      ax,.x3
-       sub      ax,.x1
-       cwde
-       movsx    ebx,bx
-       shl      eax,ROUND
-       cdq
-       idiv     ebx
- ;      mov      .dx13,eax
-       push      eax
-       mov     ax,.z3
-       sub     ax,.z1
-       cwde
-       shl     eax,CATMULL_SHIFT
-       cdq
-       idiv    ebx
-  ;    mov    .dz13,eax
-       push    eax
-if Ext>=SSE
-       sub       esp,16
-       cvtsi2ss  xmm3,ebx            ;rcps
-  ;     mov       eax,255
-       cvtsi2ss  xmm4,[i255d]
-       divss     xmm3,xmm4
-       rcpss     xmm3,xmm3
-    ;   mulss     xmm3,xmm4
-       shufps    xmm3,xmm3,0
-       movd      mm0,[.b_x1]
-       movd      mm1,[.b_x3]
-       movd      mm2,[.e_x1]
-       movd      mm3,[.e_x3]
-       pxor      mm4,mm4
-       punpcklwd mm0,mm4
-       punpcklwd mm1,mm4
-       punpcklwd mm2,mm4
-       punpcklwd mm3,mm4
-       psubd      mm1,mm0
-       psubd      mm3,mm2
-     ;  cvtpi2ps  xmm0,mm0
-     ;  movlhps   xmm0,xmm0
-     ;  cvtpi2ps  xmm0,mm2
-       cvtpi2ps  xmm1,mm1
-       movlhps   xmm1,xmm1
-       cvtpi2ps  xmm1,mm3
-    ;   subps     xmm1,xmm0
-;       cvtpi2ps  xmm0,mm3
-  ;     divps     xmm1,xmm3
-       mulps     xmm1,xmm3
-       shufps    xmm1,xmm1,10110001b
-       cvtps2pi  mm0,xmm1          ; mm0 -> 2 delta dwords
-       movhlps   xmm1,xmm1
-       cvtps2pi  mm1,xmm1
-       movq      .dey13,mm0
-       movq      .dby13,mm1
-else
-       mov      ax,word[.b_x3]
-       sub      ax,word[.b_x1]
-       cwde
-       shl      eax,ROUND
-       cdq
-       idiv     ebx
- ;      mov      .dbx13,eax
-       push      eax
-       mov      ax,word[.b_y3]
-       sub      ax,word[.b_y1]
-       cwde
-       shl      eax,ROUND
-       cdq
-       idiv     ebx
- ;      mov      .dby13,eax
-       push      eax
-       mov      ax,word[.e_x3]
-       sub      ax,word[.e_x1]
-       cwde
-       shl      eax,ROUND
-       cdq
-       idiv     ebx
- ;      mov      .dex13,eax
-       push      eax
-       mov      ax,word[.e_y3]
-       sub      ax,word[.e_y1]
-       cwde
-       shl      eax,ROUND
-       cdq
-       idiv     ebx
- ;      mov      .dey13,eax
-       push      eax
-end if
-   .bt_dx13_done:
-       mov      bx,.y3       ; calc delta23
-       sub      bx,.y2
-       jnz      .bt_dx23_make
-       mov      ecx,6
-       xor      edx,edx
-     @@:
-       push     edx   ;dword 0
-       loop     @b
-       jmp      .bt_dx23_done
- .bt_dx23_make:
-       mov      ax,.x3
-       sub      ax,.x2
-       cwde
-       movsx    ebx,bx
-       shl      eax,ROUND
-       cdq
-       idiv     ebx
- ;      mov      .dx23,eax
-       push      eax
-       mov     ax,.z3
-       sub     ax,.z2
-       cwde
-       shl     eax,CATMULL_SHIFT
-       cdq
-       idiv    ebx
-      ; mov     .dz23,eax
-       push    eax
-      ;  sub     esp,40
-if Ext>=SSE
-       sub       esp,16
-       cvtsi2ss  xmm3,ebx            ;rcps
-   ;    mov       eax,255
-       cvtsi2ss  xmm4,[i255d]
-       divss     xmm3,xmm4
-       rcpss     xmm3,xmm3
-    ;   mulss     xmm3,xmm4
-       shufps    xmm3,xmm3,0
-       movd      mm0,[.b_x2]
-       movd      mm1,[.b_x3]
-       movd      mm2,[.e_x2]
-       movd      mm3,[.e_x3]
-       pxor      mm4,mm4
-       punpcklwd mm0,mm4
-       punpcklwd mm1,mm4
-       punpcklwd mm2,mm4
-       punpcklwd mm3,mm4
-       psubd      mm1,mm0
-       psubd      mm3,mm2
-;       cvtpi2ps  xmm0,mm0
-;       movlhps   xmm0,xmm0
-;       cvtpi2ps  xmm0,mm2
-       cvtpi2ps  xmm1,mm1
-       movlhps   xmm1,xmm1
-       cvtpi2ps  xmm1,mm3
-;       subps     xmm1,xmm0
-;       cvtpi2ps  xmm0,mm3
-  ;     divps     xmm1,xmm3
-       mulps     xmm1,xmm3
-       shufps    xmm1,xmm1,10110001b
-       cvtps2pi  mm0,xmm1          ; mm0 -> 2 delta dwords
-       movhlps   xmm1,xmm1
-       cvtps2pi  mm1,xmm1
-       movq      .dey23,mm0
-       movq      .dby23,mm1
-else
-       mov      ax,word[.b_x3]
-       sub      ax,word[.b_x2]
-       cwde
-       shl      eax,ROUND
-       cdq
-       idiv     ebx
- ;      mov      .dbx23,eax
-       push      eax
-       mov      ax,word[.b_y3]
-       sub      ax,word[.b_y2]
-       cwde
-       shl      eax,ROUND
-       cdq
-       idiv     ebx
- ;      mov      .dby23,eax
-       push      eax
-       mov      ax,word[.e_x3]
-       sub      ax,word[.e_x2]
-       cwde
-       shl      eax,ROUND
-       cdq
-       idiv     ebx
- ;      mov      .dex23,eax
-       push      eax
-       mov      ax,word[.e_y3]
-       sub      ax,word[.e_y2]
-       cwde
-       shl      eax,ROUND
-       cdq
-       idiv     ebx
- ;      mov      .dey23,eax
-       push      eax
-end if
-   .bt_dx23_done:
-       sub      esp,48
-       movsx    eax,.x1
-       shl      eax,ROUND
-       mov      .cx1,eax
-       mov      .cx2,eax
-     ;  push     eax
-     ;  push     eax
-       movsx    eax,word[.b_x1]
-       shl      eax,ROUND
-       mov      .cbx1,eax
-       mov      .cbx2,eax
-    ;   push     eax
-    ;   push     eax
-       movsx    eax,word[.b_y1]
-       shl      eax,ROUND
-       mov      .cby1,eax
-       mov      .cby2,eax
-    ;   push     eax
-    ;   push     eax
-       movsx    eax,word[.e_x1]
-       shl      eax,ROUND
-       mov      .cex1,eax
-       mov      .cex2,eax
-    ;   push     eax
-    ;   push     eax
-       movsx    eax,word[.e_y1]
-       shl      eax,ROUND
-       mov      .cey1,eax
-       mov      .cey2,eax
-    ;   push     eax
-    ;   push     eax
-       movsx    eax,.z1
-       shl      eax,CATMULL_SHIFT
-       mov      .cz1,eax
-       mov      .cz2,eax
-     ;  push     eax
-     ;  push     eax
-       movsx    ecx,.y1
-       cmp      cx,.y2
-       jge      .loop12_done
-  .loop12:
-       call     .call_bump_line
-if Ext >= SSE2
-       movups   xmm0,.cey2
-       movups   xmm1,.cey1
-       movups   xmm2,.dey12
-       movups   xmm3,.dey13
-       paddd    xmm0,xmm2
-       paddd    xmm1,xmm3
-       movups   .cey2,xmm0
-       movups   .cey1,xmm1
-       movq     mm4,.cz1
-       movq     mm5,.cz2
-       paddd    mm4,.dz13
-       paddd    mm5,.dz12
-       movq     .cz1,mm4
-       movq     .cz2,mm5
-end if
-if (Ext = MMX) | (Ext = SSE)
-       movq     mm0,.cby2
-       movq     mm1,.cby1
-       movq     mm2,.cey2
-       movq     mm3,.cey1
-       movq     mm4,.cz1
-       movq     mm5,.cz2
-       paddd    mm0,.dby12
-       paddd    mm1,.dby13
-       paddd    mm2,.dey12
-       paddd    mm3,.dey13
-       paddd    mm4,.dz13
-       paddd    mm5,.dz12
-       movq     .cby2,mm0
-       movq     .cby1,mm1
-       movq     .cey1,mm3
-       movq     .cey2,mm2
-       movq     .cz1,mm4
-       movq     .cz2,mm5
-else if Ext = NON
-       mov      edx,.dbx13
-       add      .cbx1,edx
-       mov      eax,.dbx12
-       add      .cbx2,eax
-       mov      ebx,.dby13
-       add      .cby1,ebx
-       mov      edx,.dby12
-       add      .cby2,edx
-       mov      eax,.dex13
-       add      .cex1,eax
-       mov      ebx,.dex12
-       add      .cex2,ebx
-       mov      edx,.dey13
-       add      .cey1,edx
-       mov      eax,.dey12
-       add      .cey2,eax
-       mov      eax,.dx13
-       add      .cx1,eax
-       mov      ebx,.dx12
-       add      .cx2,ebx
-       mov      ebx,.dz13
-       add      .cz1,ebx
-       mov      edx,.dz12
-       add      .cz2,edx
-end if
-       inc      ecx
-       cmp      cx,.y2
-       jl       .loop12
-    .loop12_done:
-       movsx    ecx,.y2
-       cmp      cx,.y3
-       jge      .loop23_done
-       movsx    eax,.z2
-       shl      eax,CATMULL_SHIFT
-       mov      .cz2,eax
-       movsx    eax,.x2
-       shl      eax,ROUND
-       mov      .cx2,eax
-       movzx    eax,word[.b_x2]
-       shl      eax,ROUND
-       mov      .cbx2,eax
-       movzx    eax,word[.b_y2]
-       shl      eax,ROUND
-       mov      .cby2,eax
-       movzx    eax,word[.e_x2]
-       shl      eax,ROUND
-       mov      .cex2,eax
-       movzx    eax,word[.e_y2]
-       shl      eax,ROUND
-       mov      .cey2,eax
-     .loop23:
-       call     .call_bump_line
-if Ext >=  SSE2
-       movups   xmm0,.cey2
-       movups   xmm1,.cey1
-       movups   xmm2,.dey23
-       movups   xmm3,.dey13
-       paddd    xmm0,xmm2
-       paddd    xmm1,xmm3
-       movups   .cey2,xmm0
-       movups   .cey1,xmm1
-       movq     mm4,.cz1
-       movq     mm5,.cz2
-       paddd    mm4,.dz13
-       paddd    mm5,.dz23
-       movq     .cz1,mm4
-       movq     .cz2,mm5
-end if
-if (Ext = MMX) | (Ext = SSE)
-       movq     mm0,.cby2
-       movq     mm1,.cby1
-       movq     mm2,.cey2
-       movq     mm3,.cey1
-       movq     mm4,.cz1
-       movq     mm5,.cz2
-       paddd    mm0,.dby23
-       paddd    mm1,.dby13
-       paddd    mm2,.dey23
-       paddd    mm3,.dey13
-       paddd    mm4,.dz13
-       paddd    mm5,.dz23
-       movq     .cby2,mm0
-       movq     .cby1,mm1
-       movq     .cey1,mm3
-       movq     .cey2,mm2
-       movq     .cz1,mm4
-       movq     .cz2,mm5
-else if Ext = NON
-       mov      eax,.dx13
-       add      .cx1,eax
-       mov      ebx,.dx23
-       add      .cx2,ebx
-       mov      edx,.dbx13
-       add      .cbx1,edx
-       mov      eax,.dbx23
-       add      .cbx2,eax
-       mov      ebx,.dby13
-       add      .cby1,ebx
-       mov      edx,.dby23
-       add      .cby2,edx
-       mov      eax,.dex13
-       add      .cex1,eax
-       mov      ebx,.dex23
-       add      .cex2,ebx
-       mov      edx,.dey13
-       add      .cey1,edx
-       mov      eax,.dey23
-       add      .cey2,eax
-       mov      ebx,.dz13
-       add      .cz1,ebx
-       mov      edx,.dz23
-       add      .cz2,edx
-end if
-       inc      ecx
-       cmp      cx,.y3
-       jl       .loop23
-    .loop23_done:
-       mov      esp,ebp
-ret   34
-.call_bump_line:
-      ; push     ebp
-      ; push     ecx
-       pushad
-       push     dword .cz1
-       push     dword .cz2
-       push     .z_buff
-       push     .t_emap
-       push     .t_bmap
-       push     dword .cey2
-       push     .cex2
-       push     dword .cby2
-       push     .cbx2
-       push     dword .cey1
-       push     .cex1
-       push     dword .cby1
-       push     .cbx1
-       push     ecx
-       mov      eax,.cx1
-       sar      eax,ROUND
-       mov      ebx,.cx2
-       sar      ebx,ROUND
-       call     bump_line_z
-       popad
-ret
-bump_line_z:
-;--------------in: eax - x1
-;--------------    ebx - x2
-;--------------    edi - pointer to screen buffer
-;stack - another parameters :
-.y      equ dword [ebp+4]
-.bx1    equ       [ebp+8]   ;   ---
-.by1    equ dword [ebp+12]  ;       |
-.ex1    equ       [ebp+16]  ;       |
-.ey1    equ dword [ebp+20]  ;       |>   bump and env coords
-.bx2    equ       [ebp+24]  ;       |>   shifted shl ROUND
-.by2    equ dword [ebp+28]  ;       |
-.ex2    equ       [ebp+32]  ;       |
-.ey2    equ dword [ebp+36]  ;   ---
-.bmap   equ dword [ebp+40]
-.emap   equ dword [ebp+44]
-.z_buff equ dword [ebp+48]
-.z2     equ dword [ebp+52]  ;   -- |>   z coords shifted
-.z1     equ dword [ebp+56]  ;   --       shl  CATMULL_SHIFT
-.x1     equ dword [ebp-4]
-.x2     equ dword [ebp-8]
-.dbx    equ dword [ebp-12]
-.dby    equ       [ebp-16]
-.dex    equ dword [ebp-20]
-.dey    equ       [ebp-24]
-.dz     equ dword [ebp-28]
-.cbx    equ dword [ebp-32]
-.cby    equ       [ebp-36]
-.cex    equ dword [ebp-40]
-.cey    equ       [ebp-44]
-.cz     equ dword [ebp-48]
-.czbuff equ dword [ebp-52]
-.temp1  equ        ebp-60
-.temp2  equ        ebp-68
-.temp3  equ        ebp-76
-.temp4  equ        ebp-84
-.temp5  equ        ebp-92
-        mov     ebp,esp
-        mov     ecx,.y
-        or      ecx,ecx
-        jl      .bl_end
-  ;      mov     dx,[size_x_var]
-  ;      dec     dx
-        cmp     cx,[size_y_var]  ;SIZE_Y
-        jge     .bl_end
-        cmp     eax,ebx
-        jl      .bl_ok
-        je      .bl_end
-        xchg    eax,ebx
-if Ext=NON
-        mov     edx,.bx1
-        xchg    edx,.bx2
-        mov     .bx1,edx
-        mov     edx,.by1
-        xchg    edx,.by2
-        mov     .by1,edx
-        mov     edx,.ex1
-        xchg    edx,.ex2
-        mov     .ex1,edx
-        mov     edx,.ey1
-        xchg    edx,.ey2
-        mov     .ey1,edx
-end if
-if Ext = MMX
-        movq    mm0,.bx1
-        movq    mm1,.ex1
-        movq    mm2,.bx2
-        movq    mm3,.ex2
-        movq    .bx2,mm0
-        movq    .ex2,mm1
-        movq    .bx1,mm2
-        movq    .ex1,mm3
-end if
-if Ext >= SSE
-        movups  xmm0,.bx1
-        movups  xmm1,.bx2
-        movups  .bx2,xmm0
-        movups  .bx1,xmm1
-end if
-        mov     edx,.z1
-        xchg    edx,.z2
-        mov     .z1,edx
-  .bl_ok:
-        push    eax
-        push    ebx           ;store x1, x2
-        movzx   edx,word[size_x_var]
-        dec     edx
-        cmp     .x1,edx  ;SIZE_X
-        jge     .bl_end
-        cmp     .x2,0
-        jle     .bl_end
-        mov     ebx,.x2
-        sub     ebx,.x1
-if  Ext >= SSE
-       sub       esp,16
-       cvtsi2ss  xmm3,ebx            ;rcps
-       shufps    xmm3,xmm3,0
-       cvtpi2ps  xmm0,.bx1 ;mm0
-       movlhps   xmm0,xmm0
-       cvtpi2ps  xmm0,.ex1 ;mm2
-       cvtpi2ps  xmm1,.bx2 ;mm1
-       movlhps   xmm1,xmm1
-       cvtpi2ps  xmm1,.ex2 ;mm3
-       subps     xmm1,xmm0
-       divps     xmm1,xmm3
-       shufps    xmm1,xmm1,10110001b
-       cvtps2pi  mm0,xmm1          ; mm0 -> 2 delta dwords
-       movhlps   xmm1,xmm1
-       cvtps2pi  mm1,xmm1
-       movq      .dey,mm0
-       movq      .dby,mm1
-else
-        mov     eax,.bx2       ; calc .dbx
-        sub     eax,.bx1
-        cdq
-        idiv    ebx
-        push    eax
-        mov     eax,.by2       ; calc .dby
-        sub     eax,.by1
-        cdq
-        idiv    ebx
-        push    eax
-        mov     eax,.ex2       ; calc .dex
-        sub     eax,.ex1
-        cdq
-        idiv    ebx
-        push    eax
-        mov     eax,.ey2       ; calc .dey
-        sub     eax,.ey1
-        cdq
-        idiv    ebx
-        push    eax
-end if
-        mov     eax,.z2        ; calc .dz
-        sub     eax,.z1
-        cdq
-        idiv    ebx
-        push    eax
-        cmp     .x1,0         ; set correctly begin variable
-        jge     @f            ; CLIPPING ON FUNCTION
-                              ; cutting triangle exceedes screen
-        mov     ebx,.x1
-        neg     ebx
-        imul    ebx           ; eax = .dz * abs(.x1)
-        add     .z1,eax
-        mov     .x1,0
-        mov     eax,.dbx
-        imul    ebx
-        add    .bx1,eax
-        mov     eax,.dby
-        imul    ebx
-        add     .by1,eax
-        mov     eax,.dex
-        imul    ebx
-        add     .ex1,eax
-        mov     eax,.dey
-        imul    ebx
-        add     .ey1,eax
-      @@:
-        movzx   edx,word[size_x_var]
-        dec     edx
-        cmp     .x2,edx  ;SIZE_X
-        jl      @f
-        mov     .x2,edx ;SIZE_X
-      @@:
-        movzx   eax,word[size_x_var]  ;SIZE_X       ;calc memory begin in buffers
-        mov     ebx,.y
-        mul     ebx
-        mov     ebx,.x1
-        add     eax,ebx
-        mov     ebx,eax
-        lea     eax,[eax*3]
-        add     edi,eax
-        mov     esi,.z_buff       ; z-buffer filled with dd variables
-        shl     ebx,2
-        add     esi,ebx
-        mov     ecx,.x2
-        sub     ecx,.x1
-        ; init current variables
-        push    dword .bx1
-        push    .by1
-        push    dword .ex1
-        push    .ey1
-        push    .z1              ; current z shl CATMULL_SHIFT
-        push    esi
-;if Ext = SSE2
-;        movups  xmm1,.dey
-;end if
-if Ext>=MMX
-        movq    mm0,.cby
-        movq    mm1,.cey
-        movq    mm2,.dby
-        movq    mm3,.dey
-end if
-if Ext >= SSE2
-        mov    eax,TEXTURE_SIZE
-        movd   xmm1,eax
-        shufps xmm1,xmm1,0
-        push   dword  TEX_X
-        push   dword  -TEX_X
-        push   dword  1
-        push   dword  -1
-        movups xmm2,[esp]
-        movd   xmm3,.bmap
-        shufps xmm3,xmm3,0
-end if
-;align 16
-     .draw:
-    ; if TEX = SHIFTING   ;bump drawing only in shifting mode
-        mov     esi,.czbuff      ; .czbuff current address in buffer
-        mov     ebx,.cz          ; .cz - cur z position
-        cmp     ebx,dword[esi]
-        jge     .skip
-if Ext>=MMX
-        movq    mm6,mm0
-        psrld   mm6,ROUND
-        movd    eax,mm6
-        psrlq   mm6,32
-        movd    esi,mm6
-else
-        mov     eax,.cby
-        sar     eax,ROUND
-        mov     esi,.cbx
-        sar     esi,ROUND
-end if
-        shl     eax,TEX_SHIFT   ;-
-        add     esi,eax         ;-  ; esi - current bump map index
-if Ext = SSE2
-        movd    xmm0,esi
-        shufps  xmm0,xmm0,0
-        paddd   xmm0,xmm2
-        pand    xmm0,xmm1
-        paddd   xmm0,xmm3
-        movd    ebx,xmm0
-        movzx   eax,byte[ebx]
-;
-;        shufps  xmm0,xmm0,11100001b
-        psrldq  xmm0,4
-        movd    ebx,xmm0
-        movzx   ebx,byte[ebx]
-        sub     eax,ebx
-;
-;        shufps  xmm0,xmm0,11111110b
-        psrldq  xmm0,4
-        movd    ebx,xmm0
-        movzx   edx, byte [ebx]
-;
-;        shufps  xmm0,xmm0,11111111b
-        psrldq  xmm0,4
-        movd    ebx,xmm0
-        movzx   ebx, byte [ebx]
-        sub     edx,ebx
-;
-else
-        mov     ebx,esi
-        dec     ebx
-        and     ebx,TEXTURE_SIZE
-        add     ebx,.bmap
-        movzx   eax,byte [ebx]
-        mov     ebx,esi
-        inc     ebx
-        and     ebx,TEXTURE_SIZE
-        add     ebx,.bmap
-        movzx   ebx,byte [ebx]
-        sub     eax,ebx
-        mov     ebx,esi
-        sub     ebx,TEX_X
-        and     ebx,TEXTURE_SIZE
-        add     ebx,.bmap
-        movzx   edx,byte [ebx]
-        mov     ebx,esi
-        add     ebx,TEX_X
-        and     ebx,TEXTURE_SIZE
-        add     ebx,.bmap
-        movzx   ebx,byte [ebx]
-        sub     edx,ebx
-end if
-     ;  eax - horizontal sub
-     ;  edx - vertical   sub
-if Ext = NON
-        mov     ebx,.cex       ;.cex - current env map X
-        sar     ebx,ROUND
-        add     eax,ebx        ; eax - modified x coord
-        mov     ebx,.cey       ;.cey - current  env map y
-        sar     ebx,ROUND
-        add     edx,ebx        ; edx - modified y coord
-else
-        movq    mm6,mm1        ; mm5 - copy of cur env coords
-        psrld   mm6,ROUND
-        movd    ebx,mm6
-        psrlq   mm6,32
-        add     eax,ebx
-        movd    ebx,mm6
-        add     edx,ebx
-end if
-        or      eax,eax
-        jl      .black
-        cmp     eax,TEX_X
-        jg      .black
-        or      edx,edx
-        jl      .black
-        cmp     edx,TEX_Y
-        jg      .black
-        shl     edx,TEX_SHIFT
-        add     edx,eax
-        lea     esi,[edx*3]
-        add     esi,.emap
-        lodsd
-        jmp     .put_pixel
-     .black:
-        xor     eax,eax
-     .put_pixel:
-        stosd
-        dec     edi
-        mov     ebx,.cz
-        mov     esi,.czbuff
-        mov     dword[esi],ebx
-        jmp     .no_skip
-     .skip:
-        add     edi,3
-     .no_skip:
-        add     .czbuff,4
-;if Ext = SSE2
-;        movups  xmm0,.cey
-;        paddd   xmm0,xmm1
-;        movups  .cey,xmm0
-;
-;end if
-if Ext >= MMX
-        paddd   mm0,mm2
-        paddd   mm1,mm3
-end if
-if Ext=NON
-        mov     eax,.dbx
-        add     .cbx,eax
-        mov     eax,.dby
-        add     .cby,eax
-        mov     eax,.dex
-        add     .cex,eax
-        mov     eax,.dey
-        add     .cey,eax
-end if
-        mov     eax,.dz
-        add     .cz,eax
-        dec     ecx
-        jnz     .draw
-   ;   end if
-  .bl_end:
-        mov     esp,ebp
-ret 56
+;SIZE_X equ 350
+;SIZE_Y equ 350
+ROUND equ 8
+;TEX_X equ 512
+;TEX_Y equ 512
+;TEXTURE_SIZE EQU (512*512)-1
+;TEX_SHIFT EQU 9
+CATMULL_SHIFT equ 8
+;TEXTURE_SIZE EQU (TEX_X * TEX_Y)-1
+;Ext = NON
+;MMX = 1
+;NON = 0
+;------- Big thanks to Majuma (www.majuma.xt.pl) for absolutely great---
+;------- DOS 13h mode demos --------------------------------------------
+;------- Procedure draws bump triangle using Catmull Z-buffer algorithm-
+;------- (Z coordinate interpolation)-----------------------------------
+bump_triangle_z:
+;------------------in - eax - x1 shl 16 + y1 -----------
+;---------------------- ebx - x2 shl 16 + y2 -----------
+;---------------------- ecx - x3 shl 16 + y3 -----------
+;---------------------- edx - pointer to bump map ------
+;---------------------- esi - pointer to environment map
+;---------------------- edi - pointer to screen buffer--
+;---------------------- stack : bump coordinates--------
+;----------------------         environment coordinates-
+;----------------------         Z position coordinates--
+;----------------------         pointer io Z buffer-----
+;-- Z-buffer - filled with coordinates as dword --------
+;-- (Z coor. as word) shl CATMULL_SHIFT ----------------
+.b_x1   equ ebp+4   ; procedure don't save registers !!!
+.b_y1   equ ebp+6   ; each coordinate as word
+.b_x2   equ ebp+8
+.b_y2   equ ebp+10
+.b_x3   equ ebp+12
+.b_y3   equ ebp+14
+.e_x1   equ ebp+16
+.e_y1   equ ebp+18
+.e_x2   equ ebp+20
+.e_y2   equ ebp+22
+.e_x3   equ ebp+24
+.e_y3   equ ebp+26
+.z1     equ word[ebp+28]
+.z2     equ word[ebp+30]
+.z3     equ word[ebp+32]
+.z_buff equ dword[ebp+34]       ; pointer to Z-buffer
+.t_bmap equ dword[ebp-4]        ; pointer to bump map
+.t_emap equ dword[ebp-8]        ; pointer to e. map
+.x1     equ word[ebp-10]
+.y1     equ word[ebp-12]
+.x2     equ word[ebp-14]
+.y2     equ word[ebp-16]
+.x3     equ word[ebp-18]
+.y3     equ word[ebp-20]
+.dx12  equ dword[ebp-24]
+.dz12  equ      [ebp-28]
+.dbx12 equ dword[ebp-32]
+.dby12 equ      [ebp-36]
+.dex12 equ dword[ebp-40]
+.dey12 equ      [ebp-44]
+.dx13  equ dword[ebp-48]
+.dz13  equ      [ebp-52]
+.dbx13 equ dword[ebp-56]
+.dby13 equ      [ebp-60]
+.dex13 equ dword[ebp-64]
+.dey13 equ      [ebp-68]
+.dx23  equ dword[ebp-72]
+.dz23  equ      [ebp-76]
+.dbx23 equ dword[ebp-80]
+.dby23 equ      [ebp-84]
+.dex23 equ dword[ebp-88]
+.dey23 equ      [ebp-92]
+.cx1   equ dword[ebp-96]   ; current variables
+.cz1   equ      [ebp-100]
+.cx2   equ dword[ebp-104]
+.cz2   equ      [ebp-108]
+.cbx1  equ dword[ebp-112]
+.cby1  equ      [ebp-116]
+.cex1  equ dword[ebp-120]
+.cey1  equ      [ebp-124]
+.cbx2  equ dword[ebp-128]
+.cby2  equ      [ebp-132]
+.cex2  equ dword[ebp-136]
+.cey2  equ      [ebp-140]
+       mov     ebp,esp
+       push    edx        ; store bump map
+       push    esi        ; store e. map
+     ; sub     esp,120
+ .sort3:                  ; sort triangle coordinates...
+       cmp     ax,bx
+       jle     .sort1
+       xchg    eax,ebx
+       mov     edx,dword[.b_x1]
+       xchg    edx,dword[.b_x2]
+       mov     dword[.b_x1],edx
+       mov     edx,dword[.e_x1]
+       xchg    edx,dword[.e_x2]
+       mov     dword[.e_x1],edx
+       mov     dx,.z1
+       xchg    dx,.z2
+       mov     .z1,dx
+ .sort1:
+       cmp      bx,cx
+       jle      .sort2
+       xchg     ebx,ecx
+       mov      edx,dword[.b_x2]
+       xchg     edx,dword[.b_x3]
+       mov      dword[.b_x2],edx
+       mov      edx,dword[.e_x2]
+       xchg     edx,dword[.e_x3]
+       mov      dword[.e_x2],edx
+       mov     dx,.z2
+       xchg    dx,.z3
+       mov     .z2,dx
+       jmp      .sort3
+ .sort2:
+       push     eax     ; store triangle coords in variables
+       push     ebx
+       push     ecx
+         mov      edx,80008000h  ; eax,ebx,ecx are ANDd together into edx which means that
+         and      edx,ebx        ; if *all* of them are negative a sign flag is raised
+         and      edx,ecx
+         and      edx,eax
+         test     edx,80008000h  ; Check both X&Y at once
+         jne      .loop23_done
+    ;   mov     edx,eax         ; eax,ebx,ecx are ORd together into edx which means that
+    ;   or      edx,ebx         ; if any *one* of them is negative a sign flag is raised
+    ;   or      edx,ecx
+    ;   test    edx,80000000h   ; Check only X
+    ;   jne     .loop23_done
+    ;   cmp     .x1,SIZE_X    ; {
+    ;   jg      .loop23_done
+    ;   cmp     .x2,SIZE_X     ; This can be optimized with effort
+    ;   jg      .loop23_done
+    ;   cmp     .x3,SIZE_X
+    ;   jg      .loop23_done    ; {
+       mov      bx,.y2       ; calc delta 12
+       sub      bx,.y1
+       jnz      .bt_dx12_make
+       mov      ecx,6
+       xor      edx,edx
+     @@:
+       push     edx   ;dword 0
+       loop     @b
+       jmp      .bt_dx12_done
+ .bt_dx12_make:
+       mov      ax,.x2
+       sub      ax,.x1
+       cwde
+       movsx    ebx,bx
+       shl      eax,ROUND
+       cdq
+       idiv     ebx
+ ;      mov      .dx12,eax
+       push      eax
+       mov     ax,.z2
+       sub     ax,.z1
+       cwde
+       shl     eax,CATMULL_SHIFT
+       cdq
+       idiv    ebx
+       push    eax
+if Ext>=SSE
+       sub       esp,16
+       cvtsi2ss  xmm3,ebx            ;rcps
+   ;    mov       eax,255
+       cvtsi2ss  xmm4,[i255d]
+       divss     xmm3,xmm4
+       rcpss     xmm3,xmm3
+    ;   mulss     xmm3,xmm4
+       shufps    xmm3,xmm3,0
+       movd      mm0,[.b_x1]
+       movd      mm1,[.b_x2]
+       movd      mm2,[.e_x1]
+       movd      mm3,[.e_x2]
+       pxor      mm4,mm4
+       punpcklwd mm0,mm4
+       punpcklwd mm1,mm4
+       punpcklwd mm2,mm4
+       punpcklwd mm3,mm4
+       psubd      mm1,mm0
+       psubd      mm3,mm2
+ ;       cvtpi2ps  xmm0,mm0
+ ;       movlhps   xmm0,xmm0
+ ;       cvtpi2ps  xmm0,mm2
+       cvtpi2ps  xmm1,mm1
+       movlhps   xmm1,xmm1
+       cvtpi2ps  xmm1,mm3
+ ;       subps     xmm1,xmm0
+;       cvtpi2ps  xmm0,mm3
+  ;     divps     xmm1,xmm3
+       mulps     xmm1,xmm3
+       shufps    xmm1,xmm1,10110001b
+       cvtps2pi  mm0,xmm1          ; mm0 -> 2 delta dwords
+       movhlps   xmm1,xmm1
+       cvtps2pi  mm1,xmm1
+       movq      .dey12,mm0
+       movq      .dby12,mm1
+else
+       mov      ax,word[.b_x2]
+       sub      ax,word[.b_x1]
+       cwde
+       shl      eax,ROUND
+       cdq
+       idiv     ebx
+ ;      mov      .dbx12,eax
+       push      eax
+       mov      ax,word[.b_y2]
+       sub      ax,word[.b_y1]
+       cwde
+       shl      eax,ROUND
+       cdq
+       idiv     ebx
+ ;      mov      .dby12,eax
+       push      eax
+       mov      ax,word[.e_x2]
+       sub      ax,word[.e_x1]
+       cwde
+       shl      eax,ROUND
+       cdq
+       idiv     ebx
+ ;      mov      .dex12,eax
+       push      eax
+       mov      ax,word[.e_y2]
+       sub      ax,word[.e_y1]
+       cwde
+       shl      eax,ROUND
+       cdq
+       idiv     ebx
+ ;      mov      .dey12,eax
+       push      eax
+end if
+   .bt_dx12_done:
+       mov      bx,.y3       ; calc delta13
+       sub      bx,.y1
+       jnz      .bt_dx13_make
+       mov      ecx,6
+       xor      edx,edx
+     @@:
+       push     edx   ;dword 0
+       loop     @b
+       jmp      .bt_dx13_done
+ .bt_dx13_make:
+       mov      ax,.x3
+       sub      ax,.x1
+       cwde
+       movsx    ebx,bx
+       shl      eax,ROUND
+       cdq
+       idiv     ebx
+ ;      mov      .dx13,eax
+       push      eax
+       mov     ax,.z3
+       sub     ax,.z1
+       cwde
+       shl     eax,CATMULL_SHIFT
+       cdq
+       idiv    ebx
+  ;    mov    .dz13,eax
+       push    eax
+if Ext>=SSE
+       sub       esp,16
+       cvtsi2ss  xmm3,ebx            ;rcps
+  ;     mov       eax,255
+       cvtsi2ss  xmm4,[i255d]
+       divss     xmm3,xmm4
+       rcpss     xmm3,xmm3
+    ;   mulss     xmm3,xmm4
+       shufps    xmm3,xmm3,0
+       movd      mm0,[.b_x1]
+       movd      mm1,[.b_x3]
+       movd      mm2,[.e_x1]
+       movd      mm3,[.e_x3]
+       pxor      mm4,mm4
+       punpcklwd mm0,mm4
+       punpcklwd mm1,mm4
+       punpcklwd mm2,mm4
+       punpcklwd mm3,mm4
+       psubd      mm1,mm0
+       psubd      mm3,mm2
+     ;  cvtpi2ps  xmm0,mm0
+     ;  movlhps   xmm0,xmm0
+     ;  cvtpi2ps  xmm0,mm2
+       cvtpi2ps  xmm1,mm1
+       movlhps   xmm1,xmm1
+       cvtpi2ps  xmm1,mm3
+    ;   subps     xmm1,xmm0
+;       cvtpi2ps  xmm0,mm3
+  ;     divps     xmm1,xmm3
+       mulps     xmm1,xmm3
+       shufps    xmm1,xmm1,10110001b
+       cvtps2pi  mm0,xmm1          ; mm0 -> 2 delta dwords
+       movhlps   xmm1,xmm1
+       cvtps2pi  mm1,xmm1
+       movq      .dey13,mm0
+       movq      .dby13,mm1
+else
+       mov      ax,word[.b_x3]
+       sub      ax,word[.b_x1]
+       cwde
+       shl      eax,ROUND
+       cdq
+       idiv     ebx
+ ;      mov      .dbx13,eax
+       push      eax
+       mov      ax,word[.b_y3]
+       sub      ax,word[.b_y1]
+       cwde
+       shl      eax,ROUND
+       cdq
+       idiv     ebx
+ ;      mov      .dby13,eax
+       push      eax
+       mov      ax,word[.e_x3]
+       sub      ax,word[.e_x1]
+       cwde
+       shl      eax,ROUND
+       cdq
+       idiv     ebx
+ ;      mov      .dex13,eax
+       push      eax
+       mov      ax,word[.e_y3]
+       sub      ax,word[.e_y1]
+       cwde
+       shl      eax,ROUND
+       cdq
+       idiv     ebx
+ ;      mov      .dey13,eax
+       push      eax
+end if
+   .bt_dx13_done:
+       mov      bx,.y3       ; calc delta23
+       sub      bx,.y2
+       jnz      .bt_dx23_make
+       mov      ecx,6
+       xor      edx,edx
+     @@:
+       push     edx   ;dword 0
+       loop     @b
+       jmp      .bt_dx23_done
+ .bt_dx23_make:
+       mov      ax,.x3
+       sub      ax,.x2
+       cwde
+       movsx    ebx,bx
+       shl      eax,ROUND
+       cdq
+       idiv     ebx
+ ;      mov      .dx23,eax
+       push      eax
+       mov     ax,.z3
+       sub     ax,.z2
+       cwde
+       shl     eax,CATMULL_SHIFT
+       cdq
+       idiv    ebx
+      ; mov     .dz23,eax
+       push    eax
+      ;  sub     esp,40
+if Ext>=SSE
+       sub       esp,16
+       cvtsi2ss  xmm3,ebx            ;rcps
+   ;    mov       eax,255
+       cvtsi2ss  xmm4,[i255d]
+       divss     xmm3,xmm4
+       rcpss     xmm3,xmm3
+    ;   mulss     xmm3,xmm4
+       shufps    xmm3,xmm3,0
+       movd      mm0,[.b_x2]
+       movd      mm1,[.b_x3]
+       movd      mm2,[.e_x2]
+       movd      mm3,[.e_x3]
+       pxor      mm4,mm4
+       punpcklwd mm0,mm4
+       punpcklwd mm1,mm4
+       punpcklwd mm2,mm4
+       punpcklwd mm3,mm4
+       psubd      mm1,mm0
+       psubd      mm3,mm2
+;       cvtpi2ps  xmm0,mm0
+;       movlhps   xmm0,xmm0
+;       cvtpi2ps  xmm0,mm2
+       cvtpi2ps  xmm1,mm1
+       movlhps   xmm1,xmm1
+       cvtpi2ps  xmm1,mm3
+;       subps     xmm1,xmm0
+;       cvtpi2ps  xmm0,mm3
+  ;     divps     xmm1,xmm3
+       mulps     xmm1,xmm3
+       shufps    xmm1,xmm1,10110001b
+       cvtps2pi  mm0,xmm1          ; mm0 -> 2 delta dwords
+       movhlps   xmm1,xmm1
+       cvtps2pi  mm1,xmm1
+       movq      .dey23,mm0
+       movq      .dby23,mm1
+else
+       mov      ax,word[.b_x3]
+       sub      ax,word[.b_x2]
+       cwde
+       shl      eax,ROUND
+       cdq
+       idiv     ebx
+ ;      mov      .dbx23,eax
+       push      eax
+       mov      ax,word[.b_y3]
+       sub      ax,word[.b_y2]
+       cwde
+       shl      eax,ROUND
+       cdq
+       idiv     ebx
+ ;      mov      .dby23,eax
+       push      eax
+       mov      ax,word[.e_x3]
+       sub      ax,word[.e_x2]
+       cwde
+       shl      eax,ROUND
+       cdq
+       idiv     ebx
+ ;      mov      .dex23,eax
+       push      eax
+       mov      ax,word[.e_y3]
+       sub      ax,word[.e_y2]
+       cwde
+       shl      eax,ROUND
+       cdq
+       idiv     ebx
+ ;      mov      .dey23,eax
+       push      eax
+end if
+   .bt_dx23_done:
+       sub      esp,48
+       movsx    eax,.x1
+       shl      eax,ROUND
+       mov      .cx1,eax
+       mov      .cx2,eax
+     ;  push     eax
+     ;  push     eax
+       movsx    eax,word[.b_x1]
+       shl      eax,ROUND
+       mov      .cbx1,eax
+       mov      .cbx2,eax
+    ;   push     eax
+    ;   push     eax
+       movsx    eax,word[.b_y1]
+       shl      eax,ROUND
+       mov      .cby1,eax
+       mov      .cby2,eax
+    ;   push     eax
+    ;   push     eax
+       movsx    eax,word[.e_x1]
+       shl      eax,ROUND
+       mov      .cex1,eax
+       mov      .cex2,eax
+    ;   push     eax
+    ;   push     eax
+       movsx    eax,word[.e_y1]
+       shl      eax,ROUND
+       mov      .cey1,eax
+       mov      .cey2,eax
+    ;   push     eax
+    ;   push     eax
+       movsx    eax,.z1
+       shl      eax,CATMULL_SHIFT
+       mov      .cz1,eax
+       mov      .cz2,eax
+     ;  push     eax
+     ;  push     eax
+       movsx    ecx,.y1
+       cmp      cx,.y2
+       jge      .loop12_done
+  .loop12:
+       call     .call_bump_line
+if Ext >= SSE2
+       movups   xmm0,.cey2
+       movups   xmm1,.cey1
+       movups   xmm2,.dey12
+       movups   xmm3,.dey13
+       paddd    xmm0,xmm2
+       paddd    xmm1,xmm3
+       movups   .cey2,xmm0
+       movups   .cey1,xmm1
+       movq     mm4,.cz1
+       movq     mm5,.cz2
+       paddd    mm4,.dz13
+       paddd    mm5,.dz12
+       movq     .cz1,mm4
+       movq     .cz2,mm5
+end if
+if (Ext = MMX) | (Ext = SSE)
+       movq     mm0,.cby2
+       movq     mm1,.cby1
+       movq     mm2,.cey2
+       movq     mm3,.cey1
+       movq     mm4,.cz1
+       movq     mm5,.cz2
+       paddd    mm0,.dby12
+       paddd    mm1,.dby13
+       paddd    mm2,.dey12
+       paddd    mm3,.dey13
+       paddd    mm4,.dz13
+       paddd    mm5,.dz12
+       movq     .cby2,mm0
+       movq     .cby1,mm1
+       movq     .cey1,mm3
+       movq     .cey2,mm2
+       movq     .cz1,mm4
+       movq     .cz2,mm5
+else if Ext = NON
+       mov      edx,.dbx13
+       add      .cbx1,edx
+       mov      eax,.dbx12
+       add      .cbx2,eax
+       mov      ebx,.dby13
+       add      .cby1,ebx
+       mov      edx,.dby12
+       add      .cby2,edx
+       mov      eax,.dex13
+       add      .cex1,eax
+       mov      ebx,.dex12
+       add      .cex2,ebx
+       mov      edx,.dey13
+       add      .cey1,edx
+       mov      eax,.dey12
+       add      .cey2,eax
+       mov      eax,.dx13
+       add      .cx1,eax
+       mov      ebx,.dx12
+       add      .cx2,ebx
+       mov      ebx,.dz13
+       add      .cz1,ebx
+       mov      edx,.dz12
+       add      .cz2,edx
+end if
+       inc      ecx
+       cmp      cx,.y2
+       jl       .loop12
+    .loop12_done:
+       movsx    ecx,.y2
+       cmp      cx,.y3
+       jge      .loop23_done
+       movsx    eax,.z2
+       shl      eax,CATMULL_SHIFT
+       mov      .cz2,eax
+       movsx    eax,.x2
+       shl      eax,ROUND
+       mov      .cx2,eax
+       movzx    eax,word[.b_x2]
+       shl      eax,ROUND
+       mov      .cbx2,eax
+       movzx    eax,word[.b_y2]
+       shl      eax,ROUND
+       mov      .cby2,eax
+       movzx    eax,word[.e_x2]
+       shl      eax,ROUND
+       mov      .cex2,eax
+       movzx    eax,word[.e_y2]
+       shl      eax,ROUND
+       mov      .cey2,eax
+     .loop23:
+       call     .call_bump_line
+if Ext >=  SSE2
+       movups   xmm0,.cey2
+       movups   xmm1,.cey1
+       movups   xmm2,.dey23
+       movups   xmm3,.dey13
+       paddd    xmm0,xmm2
+       paddd    xmm1,xmm3
+       movups   .cey2,xmm0
+       movups   .cey1,xmm1
+       movq     mm4,.cz1
+       movq     mm5,.cz2
+       paddd    mm4,.dz13
+       paddd    mm5,.dz23
+       movq     .cz1,mm4
+       movq     .cz2,mm5
+end if
+if (Ext = MMX) | (Ext = SSE)
+       movq     mm0,.cby2
+       movq     mm1,.cby1
+       movq     mm2,.cey2
+       movq     mm3,.cey1
+       movq     mm4,.cz1
+       movq     mm5,.cz2
+       paddd    mm0,.dby23
+       paddd    mm1,.dby13
+       paddd    mm2,.dey23
+       paddd    mm3,.dey13
+       paddd    mm4,.dz13
+       paddd    mm5,.dz23
+       movq     .cby2,mm0
+       movq     .cby1,mm1
+       movq     .cey1,mm3
+       movq     .cey2,mm2
+       movq     .cz1,mm4
+       movq     .cz2,mm5
+else if Ext = NON
+       mov      eax,.dx13
+       add      .cx1,eax
+       mov      ebx,.dx23
+       add      .cx2,ebx
+       mov      edx,.dbx13
+       add      .cbx1,edx
+       mov      eax,.dbx23
+       add      .cbx2,eax
+       mov      ebx,.dby13
+       add      .cby1,ebx
+       mov      edx,.dby23
+       add      .cby2,edx
+       mov      eax,.dex13
+       add      .cex1,eax
+       mov      ebx,.dex23
+       add      .cex2,ebx
+       mov      edx,.dey13
+       add      .cey1,edx
+       mov      eax,.dey23
+       add      .cey2,eax
+       mov      ebx,.dz13
+       add      .cz1,ebx
+       mov      edx,.dz23
+       add      .cz2,edx
+end if
+       inc      ecx
+       cmp      cx,.y3
+       jl       .loop23
+    .loop23_done:
+       mov      esp,ebp
+ret   34
+.call_bump_line:
+      ; push     ebp
+      ; push     ecx
+       pushad
+       push     dword .cz1
+       push     dword .cz2
+       push     .z_buff
+       push     .t_emap
+       push     .t_bmap
+       push     dword .cey2
+       push     .cex2
+       push     dword .cby2
+       push     .cbx2
+       push     dword .cey1
+       push     .cex1
+       push     dword .cby1
+       push     .cbx1
+       push     ecx
+       mov      eax,.cx1
+       sar      eax,ROUND
+       mov      ebx,.cx2
+       sar      ebx,ROUND
+       call     bump_line_z
+       popad
+ret
+bump_line_z:
+;--------------in: eax - x1
+;--------------    ebx - x2
+;--------------    edi - pointer to screen buffer
+;stack - another parameters :
+.y      equ dword [ebp+4]
+.bx1    equ       [ebp+8]   ;   ---
+.by1    equ dword [ebp+12]  ;       |
+.ex1    equ       [ebp+16]  ;       |
+.ey1    equ dword [ebp+20]  ;       |>   bump and env coords
+.bx2    equ       [ebp+24]  ;       |>   shifted shl ROUND
+.by2    equ dword [ebp+28]  ;       |
+.ex2    equ       [ebp+32]  ;       |
+.ey2    equ dword [ebp+36]  ;   ---
+.bmap   equ dword [ebp+40]
+.emap   equ dword [ebp+44]
+.z_buff equ dword [ebp+48]
+.z2     equ dword [ebp+52]  ;   -- |>   z coords shifted
+.z1     equ dword [ebp+56]  ;   --       shl  CATMULL_SHIFT
+.x1     equ dword [ebp-4]
+.x2     equ dword [ebp-8]
+.dbx    equ dword [ebp-12]
+.dby    equ       [ebp-16]
+.dex    equ dword [ebp-20]
+.dey    equ       [ebp-24]
+.dz     equ dword [ebp-28]
+.cbx    equ dword [ebp-32]
+.cby    equ       [ebp-36]
+.cex    equ dword [ebp-40]
+.cey    equ       [ebp-44]
+.cz     equ dword [ebp-48]
+.czbuff equ dword [ebp-52]
+.temp1  equ        ebp-60
+.temp2  equ        ebp-68
+.temp3  equ        ebp-76
+.temp4  equ        ebp-84
+.temp5  equ        ebp-92
+        mov     ebp,esp
+        mov     ecx,.y
+        or      ecx,ecx
+        jl      .bl_end
+  ;      mov     dx,[size_x_var]
+  ;      dec     dx
+        cmp     cx,[size_y_var]  ;SIZE_Y
+        jge     .bl_end
+        cmp     eax,ebx
+        jl      .bl_ok
+        je      .bl_end
+        xchg    eax,ebx
+if Ext=NON
+        mov     edx,.bx1
+        xchg    edx,.bx2
+        mov     .bx1,edx
+        mov     edx,.by1
+        xchg    edx,.by2
+        mov     .by1,edx
+        mov     edx,.ex1
+        xchg    edx,.ex2
+        mov     .ex1,edx
+        mov     edx,.ey1
+        xchg    edx,.ey2
+        mov     .ey1,edx
+end if
+if Ext = MMX
+        movq    mm0,.bx1
+        movq    mm1,.ex1
+        movq    mm2,.bx2
+        movq    mm3,.ex2
+        movq    .bx2,mm0
+        movq    .ex2,mm1
+        movq    .bx1,mm2
+        movq    .ex1,mm3
+end if
+if Ext >= SSE
+        movups  xmm0,.bx1
+        movups  xmm1,.bx2
+        movups  .bx2,xmm0
+        movups  .bx1,xmm1
+end if
+        mov     edx,.z1
+        xchg    edx,.z2
+        mov     .z1,edx
+  .bl_ok:
+        push    eax
+        push    ebx           ;store x1, x2
+        movzx   edx,word[size_x_var]
+        dec     edx
+        cmp     .x1,edx  ;SIZE_X
+        jge     .bl_end
+        cmp     .x2,0
+        jle     .bl_end
+        mov     ebx,.x2
+        sub     ebx,.x1
+if  Ext >= SSE
+       sub       esp,16
+       cvtsi2ss  xmm3,ebx            ;rcps
+       shufps    xmm3,xmm3,0
+       cvtpi2ps  xmm0,.bx1 ;mm0
+       movlhps   xmm0,xmm0
+       cvtpi2ps  xmm0,.ex1 ;mm2
+       cvtpi2ps  xmm1,.bx2 ;mm1
+       movlhps   xmm1,xmm1
+       cvtpi2ps  xmm1,.ex2 ;mm3
+       subps     xmm1,xmm0
+       divps     xmm1,xmm3
+       shufps    xmm1,xmm1,10110001b
+       cvtps2pi  mm0,xmm1          ; mm0 -> 2 delta dwords
+       movhlps   xmm1,xmm1
+       cvtps2pi  mm1,xmm1
+       movq      .dey,mm0
+       movq      .dby,mm1
+else
+        mov     eax,.bx2       ; calc .dbx
+        sub     eax,.bx1
+        cdq
+        idiv    ebx
+        push    eax
+        mov     eax,.by2       ; calc .dby
+        sub     eax,.by1
+        cdq
+        idiv    ebx
+        push    eax
+        mov     eax,.ex2       ; calc .dex
+        sub     eax,.ex1
+        cdq
+        idiv    ebx
+        push    eax
+        mov     eax,.ey2       ; calc .dey
+        sub     eax,.ey1
+        cdq
+        idiv    ebx
+        push    eax
+end if
+        mov     eax,.z2        ; calc .dz
+        sub     eax,.z1
+        cdq
+        idiv    ebx
+        push    eax
+        cmp     .x1,0         ; set correctly begin variable
+        jge     @f            ; CLIPPING ON FUNCTION
+                              ; cutting triangle exceedes screen
+        mov     ebx,.x1
+        neg     ebx
+        imul    ebx           ; eax = .dz * abs(.x1)
+        add     .z1,eax
+        mov     .x1,0
+        mov     eax,.dbx
+        imul    ebx
+        add    .bx1,eax
+        mov     eax,.dby
+        imul    ebx
+        add     .by1,eax
+        mov     eax,.dex
+        imul    ebx
+        add     .ex1,eax
+        mov     eax,.dey
+        imul    ebx
+        add     .ey1,eax
+      @@:
+        movzx   edx,word[size_x_var]
+        dec     edx
+        cmp     .x2,edx  ;SIZE_X
+        jl      @f
+        mov     .x2,edx ;SIZE_X
+      @@:
+        movzx   eax,word[size_x_var]  ;SIZE_X       ;calc memory begin in buffers
+        mov     ebx,.y
+        mul     ebx
+        mov     ebx,.x1
+        add     eax,ebx
+        mov     ebx,eax
+        lea     eax,[eax*3]
+        add     edi,eax
+        mov     esi,.z_buff       ; z-buffer filled with dd variables
+        shl     ebx,2
+        add     esi,ebx
+        mov     ecx,.x2
+        sub     ecx,.x1
+        ; init current variables
+        push    dword .bx1
+        push    .by1
+        push    dword .ex1
+        push    .ey1
+        push    .z1              ; current z shl CATMULL_SHIFT
+        push    esi
+;if Ext = SSE2
+;        movups  xmm1,.dey
+;end if
+if Ext>=MMX
+        movq    mm0,.cby
+        movq    mm1,.cey
+        movq    mm2,.dby
+        movq    mm3,.dey
+end if
+if Ext >= SSE2
+        mov    eax,TEXTURE_SIZE
+        movd   xmm1,eax
+        shufps xmm1,xmm1,0
+        push   dword  TEX_X
+        push   dword  -TEX_X
+        push   dword  1
+        push   dword  -1
+        movups xmm2,[esp]
+        movd   xmm3,.bmap
+        shufps xmm3,xmm3,0
+end if
+;align 16
+     .draw:
+    ; if TEX = SHIFTING   ;bump drawing only in shifting mode
+        mov     esi,.czbuff      ; .czbuff current address in buffer
+        mov     ebx,.cz          ; .cz - cur z position
+        cmp     ebx,dword[esi]
+        jge     .skip
+if Ext>=MMX
+        movq    mm6,mm0
+        psrld   mm6,ROUND
+        movd    eax,mm6
+        psrlq   mm6,32
+        movd    esi,mm6
+else
+        mov     eax,.cby
+        sar     eax,ROUND
+        mov     esi,.cbx
+        sar     esi,ROUND
+end if
+        shl     eax,TEX_SHIFT   ;-
+        add     esi,eax         ;-  ; esi - current bump map index
+if Ext = SSE2
+        movd    xmm0,esi
+        shufps  xmm0,xmm0,0
+        paddd   xmm0,xmm2
+        pand    xmm0,xmm1
+        paddd   xmm0,xmm3
+        movd    ebx,xmm0
+        movzx   eax,byte[ebx]
+;
+;        shufps  xmm0,xmm0,11100001b
+        psrldq  xmm0,4
+        movd    ebx,xmm0
+        movzx   ebx,byte[ebx]
+        sub     eax,ebx
+;
+;        shufps  xmm0,xmm0,11111110b
+        psrldq  xmm0,4
+        movd    ebx,xmm0
+        movzx   edx, byte [ebx]
+;
+;        shufps  xmm0,xmm0,11111111b
+        psrldq  xmm0,4
+        movd    ebx,xmm0
+        movzx   ebx, byte [ebx]
+        sub     edx,ebx
+;
+else
+        mov     ebx,esi
+        dec     ebx
+        and     ebx,TEXTURE_SIZE
+        add     ebx,.bmap
+        movzx   eax,byte [ebx]
+        mov     ebx,esi
+        inc     ebx
+        and     ebx,TEXTURE_SIZE
+        add     ebx,.bmap
+        movzx   ebx,byte [ebx]
+        sub     eax,ebx
+        mov     ebx,esi
+        sub     ebx,TEX_X
+        and     ebx,TEXTURE_SIZE
+        add     ebx,.bmap
+        movzx   edx,byte [ebx]
+        mov     ebx,esi
+        add     ebx,TEX_X
+        and     ebx,TEXTURE_SIZE
+        add     ebx,.bmap
+        movzx   ebx,byte [ebx]
+        sub     edx,ebx
+end if
+     ;  eax - horizontal sub
+     ;  edx - vertical   sub
+if Ext = NON
+        mov     ebx,.cex       ;.cex - current env map X
+        sar     ebx,ROUND
+        add     eax,ebx        ; eax - modified x coord
+        mov     ebx,.cey       ;.cey - current  env map y
+        sar     ebx,ROUND
+        add     edx,ebx        ; edx - modified y coord
+else
+        movq    mm6,mm1        ; mm5 - copy of cur env coords
+        psrld   mm6,ROUND
+        movd    ebx,mm6
+        psrlq   mm6,32
+        add     eax,ebx
+        movd    ebx,mm6
+        add     edx,ebx
+end if
+        or      eax,eax
+        jl      .black
+        cmp     eax,TEX_X
+        jg      .black
+        or      edx,edx
+        jl      .black
+        cmp     edx,TEX_Y
+        jg      .black
+        shl     edx,TEX_SHIFT
+        add     edx,eax
+        lea     esi,[edx*3]
+        add     esi,.emap
+        lodsd
+        jmp     .put_pixel
+     .black:
+        xor     eax,eax
+     .put_pixel:
+        stosd
+        dec     edi
+        mov     ebx,.cz
+        mov     esi,.czbuff
+        mov     dword[esi],ebx
+        jmp     .no_skip
+     .skip:
+        add     edi,3
+     .no_skip:
+        add     .czbuff,4
+;if Ext = SSE2
+;        movups  xmm0,.cey
+;        paddd   xmm0,xmm1
+;        movups  .cey,xmm0
+;
+;end if
+if Ext >= MMX
+        paddd   mm0,mm2
+        paddd   mm1,mm3
+end if
+if Ext=NON
+        mov     eax,.dbx
+        add     .cbx,eax
+        mov     eax,.dby
+        add     .cby,eax
+        mov     eax,.dex
+        add     .cex,eax
+        mov     eax,.dey
+        add     .cey,eax
+end if
+        mov     eax,.dz
+        add     .cz,eax
+        dec     ecx
+        jnz     .draw
+   ;   end if
+  .bl_end:
+        mov     esp,ebp
+ret 56

 /programs/demos/view3ds/bump_tex.inc
 ,1817 → 1,1817
-;CATMULL_SHIFT equ 8
-;TEXTURE_SIZE EQU (TEX_X * TEX_Y)-1
-;ROUND equ 8
-;Ext = NON
-;MMX = 1
-;NON = 0
-;------- Big thanks to Majuma (www.majuma.xt.pl) for absolutely great---
-;------- DOS 13h mode demos --------------------------------------------
-;------- Procedure draws bump triangle with texture, I use -------------
-;--------Catmull Z-buffer algorithm- (Z coordinate interpolation)-------
-;--------I calc texture pixel by this way: col1*col2/256 ---------------
-bump_tex_triangle_z:
-;------------------in - eax - x1 shl 16 + y1 -----------
-;---------------------- ebx - x2 shl 16 + y2 -----------
-;---------------------- ecx - x3 shl 16 + y3 -----------
-;---------------------- edx - pointer to bump map-------
-;---------------------- esi - pointer to env map--------
-;---------------------- edi - pointer to screen buffer--
-;---------------------- stack : bump coordinates--------
-;----------------------         environment coordinates-
-;----------------------         Z position coordinates--
-;----------------------         pointer to Z buffer-----
-;----------------------         pointer to texture------
-;----------------------         texture coordinates-----
-;-- Z-buffer - filled with coordinates as dword --------
-;-- (Z coor. as word) shl CATMULL_SHIFT ----------------
-.b_x1   equ ebp+4   ; procedure don't save registers !!!
-.b_y1   equ ebp+6   ; each coordinate as word
-.b_x2   equ ebp+8
-.b_y2   equ ebp+10       ; b - bump map coords
-.b_x3   equ ebp+12       ; e - env map coords
-.b_y3   equ ebp+14
-.e_x1   equ ebp+16
-.e_y1   equ ebp+18
-.e_x2   equ ebp+20
-.e_y2   equ ebp+22
-.e_x3   equ ebp+24
-.e_y3   equ ebp+26
-.z1     equ word[ebp+28]
-.z2     equ word[ebp+30]
-.z3     equ word[ebp+32]
-.z_buff equ dword[ebp+34]       ; pointer to Z-buffer
-.tex_ptr equ dword[ebp+38]      ; ptr to texture
-.t_x1   equ ebp+42              ; texture coords
-.t_y1   equ ebp+44
-.t_x2   equ ebp+46
-.t_y2   equ ebp+48
-.t_x3   equ ebp+50
-.t_y3   equ ebp+52
-.t_bmap equ dword[ebp-4]        ; pointer to bump map
-.t_emap equ dword[ebp-8]        ; pointer to env map
-.x1     equ word[ebp-10]
-.y1     equ word[ebp-12]
-.x2     equ word[ebp-14]
-.y2     equ word[ebp-16]
-.x3     equ word[ebp-18]
-.y3     equ word[ebp-20]
-if 0 ;Ext <= SSE2
-.dx12   equ dword[edi-4]
-.dz12   equ      [edi-8]
-.dbx12  equ dword[edi-12]
-.dby12  equ      [edi-16]
-.dex12  equ dword[edi-20]
-.dey12  equ      [edi-24]
-.dtx12  equ dword[edi-28]
-.dty12  equ      [edi-32]
-.dx13  equ dword[ebp-52-4*1]
-.dz13  equ      [ebp-52-4*2]
-.dbx13 equ dword[ebp-52-4*3]
-.dby13 equ      [ebp-52-4*4]
-.dex13 equ dword[ebp-52-4*5]
-.dey13 equ      [ebp-52-4*6]
-.dtx13 equ dword[ebp-52-4*7]
-.dty13 equ      [ebp-52-4*8]
-.dx23  equ dword[ebp-(52+4*9)]
-.dz23  equ      [ebp-(52+4*10)]
-.dbx23 equ dword[ebp-(52+4*11)]
-.dby23 equ      [ebp-(52+4*12)]
-.dex23 equ dword[ebp-(52+4*13)]
-.dey23 equ      [ebp-(52+4*14)]
-.dtx23 equ dword[ebp-(52+4*15)]
-.dty23 equ      [ebp-(52+4*16)]
-else
-.dx12   equ dword[ebp-24]
-.dz12   equ      [ebp-28]
-.dbx12  equ dword[ebp-32]
-.dby12  equ      [ebp-36]
-.dex12  equ dword[ebp-40]
-.dey12  equ      [ebp-44]
-.dtx12  equ dword[ebp-48]
-.dty12  equ      [ebp-52]
-.dx13  equ dword[ebp-52-4*1]
-.dz13  equ      [ebp-52-4*2]
-.dbx13 equ dword[ebp-52-4*3]
-.dby13 equ      [ebp-52-4*4]
-.dex13 equ dword[ebp-52-4*5]
-.dey13 equ      [ebp-52-4*6]
-.dtx13 equ dword[ebp-52-4*7]
-.dty13 equ      [ebp-52-4*8]
-.dx23  equ dword[ebp-(52+4*9)]
-.dz23  equ      [ebp-(52+4*10)]
-.dbx23 equ dword[ebp-(52+4*11)]
-.dby23 equ      [ebp-(52+4*12)]
-.dex23 equ dword[ebp-(52+4*13)]
-.dey23 equ      [ebp-(52+4*14)]
-.dtx23 equ dword[ebp-(52+4*15)]
-.dty23 equ      [ebp-(52+4*16)]
-end if
-if Ext < SSE
-.cx1   equ dword[ebp-(52+4*17)]         ; current variables
-.cz1   equ      [ebp-(52+4*18)]
-.cx2   equ dword[ebp-(52+4*19)]
-.cz2   equ      [ebp-(52+4*20)]
-.cbx1  equ dword[ebp-(52+4*21)]
-.cby1  equ      [ebp-(52+4*22)]
-.cbx2  equ dword[ebp-(52+4*23)]
-.cby2  equ      [ebp-(52+4*24)]
-.cex1  equ dword[ebp-(52+4*25)]
-.cey1  equ      [ebp-(52+4*26)]
-.cex2  equ dword[ebp-(52+4*27)]
-.cey2  equ      [ebp-(52+4*28)]
-.ctx1  equ dword[ebp-(52+4*29)]
-.cty1  equ      [ebp-(52+4*30)]
-.ctx2  equ dword[ebp-(52+4*31)]
-.cty2  equ      [ebp-(52+4*32)]
-else
-.cx1   equ dword[ebp-(52+4*17)]         ; current variables
-.cz1   equ      [ebp-(52+4*18)]
-.cbx1  equ dword[ebp-(52+4*19)]
-.cby1  equ      [ebp-(52+4*20)]
-.cex1  equ dword[ebp-(52+4*21)]
-.cey1  equ      [ebp-(52+4*22)]
-.ctx1  equ dword[ebp-(52+4*23)]
-.cty1  equ      [ebp-(52+4*24)]
-.cx2   equ dword[ebp-(52+4*25)]
-.cz2   equ      [ebp-(52+4*26)]
-.cbx2  equ dword[ebp-(52+4*27)]
-.cby2  equ      [ebp-(52+4*28)]
-.cex2  equ dword[ebp-(52+4*29)]
-.cey2  equ      [ebp-(52+4*30)]
-.ctx2  equ dword[ebp-(52+4*31)]
-.cty2  equ      [ebp-(52+4*32)]
-end if
-       cld
-       mov     ebp,esp
-       push    edx        ; store bump map
-       push    esi        ; store e. map
-     ; sub     esp,120
- .sort3:                  ; sort triangle coordinates...
-       cmp     ax,bx
-       jle     .sort1
-       xchg    eax,ebx
-       mov     edx,dword[.b_x1]
-       xchg    edx,dword[.b_x2]
-       mov     dword[.b_x1],edx
-       mov     edx,dword[.e_x1]
-       xchg    edx,dword[.e_x2]
-       mov     dword[.e_x1],edx
-       mov     edx,dword[.t_x1]
-       xchg    edx,dword[.t_x2]
-       mov     dword[.t_x1],edx
-       mov     dx,.z1
-       xchg    dx,.z2
-       mov     .z1,dx
- .sort1:
-       cmp      bx,cx
-       jle      .sort2
-       xchg     ebx,ecx
-       mov      edx,dword[.b_x2]
-       xchg     edx,dword[.b_x3]
-       mov      dword[.b_x2],edx
-       mov      edx,dword[.e_x2]
-       xchg     edx,dword[.e_x3]
-       mov      dword[.e_x2],edx
-       mov      edx,dword[.t_x2]
-       xchg     edx,dword[.t_x3]
-       mov      dword[.t_x2],edx
-       mov     dx,.z2
-       xchg    dx,.z3
-       mov     .z2,dx
-       jmp      .sort3
- .sort2:
-       push     eax     ; store triangle coords in variables
-       push     ebx
-       push     ecx
-         mov      edx,80008000h  ; eax,ebx,ecx are ANDd together into edx which means that
-         and      edx,ebx        ; if *all* of them are negative a sign flag is raised
-         and      edx,ecx
-         and      edx,eax
-         test     edx,80008000h  ; Check both X&Y at once
-         jne      .loop23_done
-    ;   mov     edx,eax         ; eax,ebx,ecx are ORd together into edx which means that
-    ;   or      edx,ebx         ; if any *one* of them is negative a sign flag is raised
-    ;   or      edx,ecx
-    ;   test    edx,80000000h   ; Check only X
-    ;   jne     .loop23_done
-    ;   cmp     .x1,SIZE_X    ; {
-    ;   jg      .loop23_done
-    ;   cmp     .x2,SIZE_X     ; This can be optimized with effort
-    ;   jg      .loop23_done
-    ;   cmp     .x3,SIZE_X
-    ;   jg      .loop23_done    ; {
-       mov      bx,.y2       ; calc delta 12
-       sub      bx,.y1
-       jnz      .bt_dx12_make
-if 0 ;Ext >= SSE2
-       pxor     xmm0,xmm0
-       movups   .dty12,xmm0
-       movups   .dey12,xmm0
-       sub      esp,16
-else
-       mov      ecx,8
-       xor      edx,edx
-     @@:
-       push     edx   ;dword 0
-       loop     @b
-end if
-       jmp      .bt_dx12_done
- .bt_dx12_make:
-       movsx    ebx,bx
-if Ext>=SSE
-       sub       esp,32
-   ;    mov       eax,256
-       cvtsi2ss  xmm4,[i255d]
-       cvtsi2ss  xmm3,ebx ;rcps
-if 0 ;Ext >= SSE2
-       mov       edi,ebp
-       sub       edi,512
-       or        edi,0x0000000f
-end if
-       divss     xmm3,xmm4
-       shufps    xmm3,xmm3,0
-       movd      mm0,[.b_x1]
-       movd      mm1,[.b_x2]
-       movd      mm2,[.e_x1]
-       movd      mm3,[.e_x2]
-       pxor       mm4,mm4
-       punpcklwd  mm0,mm4
-       punpcklwd  mm1,mm4
-       punpcklwd  mm2,mm4
-       punpcklwd  mm3,mm4
-       psubd      mm1,mm0
-       psubd      mm3,mm2
-       cvtpi2ps  xmm1,mm1
-       movlhps   xmm1,xmm1
-       cvtpi2ps  xmm1,mm3
-       divps     xmm1,xmm3   ;xmm1--> | dby | dbx | dey | dex |
-       shufps    xmm1,xmm1,10110001b
-                             ;xmm1--> | dbx | dby | dex | dey |
-;1       movups    .dey12,xmm1
-       cvtps2pi  mm0,xmm1 ;mm0,xmm1          ; mm0 -> 2 delta dwords
-       movhlps   xmm1,xmm1
-       cvtps2pi  mm1,xmm1 ;mm1,xmm1
-       movq      .dey12,mm0
-       movq      .dby12,mm1
-;-------------
-  ;    pxor      mm0,mm0
-  ;    pxor      mm1,mm1
-   ;/   pinsrw    mm0,.z1,1
-   ;/   pinsrw    mm0,.x1,0
-   ;/   pinsrw    mm1,.z2,1
-   ;/   pinsrw    mm1,.x2,0
-       mov       ax,.z2
-       sub       ax,.z1
-       cwde
-       mov      dx,.x2
-       sub      dx,.x1
-       movsx    edx,dx
-   ;/    movd      mm1,eax
-   ;/    punpcklwd  mm0,mm4
-   ;/    punpcklwd  mm1,mm4
-  ;     cvtpi2ps   xmm1,mm1
-  ;     cvtpi2ps   xmm2,mm0
-  ;     subps      xmm1,xmm2
-   ;/   psubd      mm1,mm0
-       movd       mm2,[.t_x1]
-       movd       mm3,[.t_x2]
-       punpcklwd  mm2,mm4
-       punpcklwd  mm3,mm4
-       psubd      mm3,mm2
-   ;/  cvtpi2ps  xmm1,mm1
-       cvtsi2ss  xmm1,eax
-       movlhps   xmm1,xmm1
-       cvtsi2ss  xmm1,edx
-   ;    movss     xmm1,xmm4
-       shufps    xmm1,xmm1,00101111b
-       cvtpi2ps  xmm1,mm3
-       divps     xmm1,xmm3   ; xmm1--> | dx | dz | dty | dtx |
-       shufps    xmm1,xmm1,11100001b
-                             ; xmm1--> | dx | dz | dtx | dty |
-;1       movlps    .dty12,xmm1
-;1       movhps    .dz12,xmm1
-       cvtps2pi  mm0,xmm1    ; mm0 -> 2 delta dwords  | dtx | dty |
-       movhlps   xmm1,xmm1
-       cvtps2pi  mm1,xmm1
-       movq      .dty12,mm0
-       movq      .dz12,mm1
-;----
-;       mov       ax,.z2
-;       sub       ax,.z1
-;       cwde
-;       mov       bx,.x2
-;       sub       bx,.x1
-;       movsx     ebx,bx
-;       movd      mm1,eax
-;       psllq     mm1,32
-;       movd      mm1,ebx
-;;       push      ebx
-;;       push      eax
-;;       movq      mm1,[esp]
-;;       add       esp,8
-;;;       mov       ax,.z1
-;;;       mov       bx,.z2
-;;;       shl       eax,16
-;;;       shl       ebx,16
-;;;       mov       ax,.x1
-;;;       mov       bx,.x2
-;       movd       mm2,[.t_x1]
-;       movd       mm3,[.t_x2]
-;;       movd      mm0,eax
-;;       movd      mm1,ebx
-;       pxor       mm4,mm4
-;;       punpcklwd  mm0,mm4
-;;       punpcklwd  mm1,mm4
-;       punpcklwd  mm2,mm4
-;       punpcklwd  mm3,mm4
-;;       psubd    mm1,mm0
-;       psubd      mm3,mm2
-;       cvtpi2ps  xmm1,mm1
-;       movlhps   xmm1,xmm1
-;       cvtpi2ps  xmm1,mm3
-;       divps     xmm1,xmm3   ; xmm1--> | dz | dx | dty | dtx |
-;       shufps    xmm1,xmm1,10110001b
-                             ; xmm1--> | dx | dz | dtx | dty |
-;       cvtps2pi  mm0,xmm1    ; mm0 -> 2 delta dwords  | dtx | dty |
-;       movhlps   xmm1,xmm1
-;       cvtps2pi  mm1,xmm1    ; mm1 --> 2 delta dwords | dx | dz |
-;       movq      .dty12,mm0
-;       movq      .dz12,mm1
-else
-       mov      ax,.x2
-       sub      ax,.x1
-       cwde
-       shl      eax,ROUND
-       cdq
-       idiv     ebx
- ;     mov      .dx12,eax
-       push      eax
-       mov     ax,.z2
-       sub     ax,.z1
-       cwde
-       shl     eax,CATMULL_SHIFT
-       cdq
-       idiv    ebx
-       push    eax
-       mov      ax,word[.b_x2]
-       sub      ax,word[.b_x1]
-       cwde
-       shl      eax,ROUND
-       cdq
-       idiv     ebx
- ;     mov      .dbx12,eax
-       push      eax
-       mov      ax,word[.b_y2]
-       sub      ax,word[.b_y1]
-       cwde
-       shl      eax,ROUND
-       cdq
-       idiv     ebx
- ;     mov      .dby12,eax
-       push      eax
-       mov      ax,word[.e_x2]
-       sub      ax,word[.e_x1]
-       cwde
-       shl      eax,ROUND
-       cdq
-       idiv     ebx
- ;     mov      .dex12,eax
-       push      eax
-       mov      ax,word[.e_y2]
-       sub      ax,word[.e_y1]
-       cwde
-       shl      eax,ROUND
-       cdq
-       idiv     ebx
- ;     mov      .dey12,eax
-       push      eax
-       mov      ax,word[.t_x2]
-       sub      ax,word[.t_x1]
-       cwde
-       shl      eax,ROUND
-       cdq
-       idiv     ebx
- ;     mov      .dtx12,eax
-       push      eax
-       mov      ax,word[.t_y2]
-       sub      ax,word[.t_y1]
-       cwde
-       shl      eax,ROUND
-       cdq
-       idiv     ebx
- ;     mov      .dty12,eax
-       push      eax
-end if
-   .bt_dx12_done:
-       mov      bx,.y3       ; calc delta13
-       sub      bx,.y1
-       jnz      .bt_dx13_make
-       mov      ecx,8
-       xor      edx,edx
-     @@:
-       push     edx   ;dword 0
-       loop     @b
-       jmp      .bt_dx13_done
- .bt_dx13_make:
-       movsx    ebx,bx
-if Ext>=SSE
-       sub       esp,32
-   ;    mov       eax,256
-       cvtsi2ss  xmm4,[i255d]
-       cvtsi2ss  xmm3,ebx            ;rcps
-       divss     xmm3,xmm4
-       shufps    xmm3,xmm3,0
-       movd      mm0,[.b_x1]
-       movd      mm1,[.b_x3]
-       movd      mm2,[.e_x1]
-       movd      mm3,[.e_x3]
-       pxor       mm4,mm4
-       punpcklwd  mm0,mm4
-       punpcklwd  mm1,mm4
-       punpcklwd  mm2,mm4
-       punpcklwd  mm3,mm4
-       psubd      mm1,mm0
-       psubd      mm3,mm2
-       cvtpi2ps  xmm1,mm1
-       movlhps   xmm1,xmm1
-       cvtpi2ps  xmm1,mm3
-       divps     xmm1,xmm3   ;xmm1--> | dby | dbx | dey | dex |
-       shufps    xmm1,xmm1,10110001b
-                             ;xmm1--> | dbx | dby | dex | dey |
-;1       movups    .dey13,xmm1
-       cvtps2pi  mm0,xmm1 ;mm0,xmm1          ; mm0 -> 2 delta dwords
-       movhlps   xmm1,xmm1
-       cvtps2pi  mm1,xmm1 ;mm1,xmm1
-       movq      .dey13,mm0
-       movq      .dby13,mm1
-       mov       ax,.z3
-       sub       ax,.z1
-       cwde
-       mov      dx,.x3
-       sub      dx,.x1
-       movsx    edx,dx
-       movd       mm2,[.t_x1]
-       movd       mm3,[.t_x3]
-       punpcklwd  mm2,mm4
-       punpcklwd  mm3,mm4
-       psubd      mm3,mm2
-       cvtsi2ss  xmm1,eax
-       movlhps   xmm1,xmm1
-       cvtsi2ss  xmm1,edx
-       shufps    xmm1,xmm1,00101111b
-       cvtpi2ps  xmm1,mm3
-       divps     xmm1,xmm3   ; xmm1--> | dx | dz | dty | dtx |
-       shufps    xmm1,xmm1,11100001b
-                             ; xmm1--> | dx | dz | dtx | dty |
-;1       movlps    .dty13,xmm1
-;1       movhps    .dz13,xmm1
-       cvtps2pi  mm0,xmm1    ; mm0 -> 2 delta dwords  | dtx | dty |
-       movhlps   xmm1,xmm1
-       cvtps2pi  mm1,xmm1
-       movq      .dty13,mm0
-       movq      .dz13,mm1
-else
-       mov      ax,.x3
-       sub      ax,.x1
-       cwde
-       shl      eax,ROUND
-       cdq
-       idiv     ebx
- ;     mov      .dx13,eax
-       push      eax
-       mov     ax,.z3
-       sub     ax,.z1
-       cwde
-       shl     eax,CATMULL_SHIFT
-       cdq
-       idiv    ebx
-  ;    mov    .dz13,eax
-       push    eax
-       mov      ax,word[.b_x3]
-       sub      ax,word[.b_x1]
-       cwde
-       shl      eax,ROUND
-       cdq
-       idiv     ebx
- ;     mov      .dbx13,eax
-       push      eax
-       mov      ax,word[.b_y3]
-       sub      ax,word[.b_y1]
-       cwde
-       shl      eax,ROUND
-       cdq
-       idiv     ebx
- ;     mov      .dby13,eax
-       push      eax
-       mov      ax,word[.e_x3]
-       sub      ax,word[.e_x1]
-       cwde
-       shl      eax,ROUND
-       cdq
-       idiv     ebx
- ;     mov      .dex13,eax
-       push      eax
-       mov      ax,word[.e_y3]
-       sub      ax,word[.e_y1]
-       cwde
-       shl      eax,ROUND
-       cdq
-       idiv     ebx
- ;     mov      .dey13,eax
-       push      eax
-       mov      ax,word[.t_x3]
-       sub      ax,word[.t_x1]
-       cwde
-       shl      eax,ROUND
-       cdq
-       idiv     ebx
- ;     mov      .dtx13,eax
-       push      eax
-       mov      ax,word[.t_y3]
-       sub      ax,word[.t_y1]
-       cwde
-       shl      eax,ROUND
-       cdq
-       idiv     ebx
- ;     mov      .dty13,eax
-       push      eax
-end if
-   .bt_dx13_done:
-       mov      bx,.y3       ; calc delta23
-       sub      bx,.y2
-       jnz      .bt_dx23_make
-       mov      ecx,8
-       xor      edx,edx
-     @@:
-       push     edx   ;dword 0
-       loop     @b
-       jmp      .bt_dx23_done
- .bt_dx23_make:
-       movsx    ebx,bx
-if Ext>=SSE
-       sub       esp,32
-   ;    mov       eax,256
-       cvtsi2ss  xmm4,[i255d]
-       cvtsi2ss  xmm3,ebx            ;rcps
-       divss     xmm3,xmm4
-       shufps    xmm3,xmm3,0
-       movd      mm0,[.b_x2]
-       movd      mm1,[.b_x3]
-       movd      mm2,[.e_x2]
-       movd      mm3,[.e_x3]
-       pxor       mm4,mm4
-       punpcklwd  mm0,mm4
-       punpcklwd  mm1,mm4
-       punpcklwd  mm2,mm4
-       punpcklwd  mm3,mm4
-       psubd      mm1,mm0
-       psubd      mm3,mm2
-       cvtpi2ps  xmm1,mm1
-       movlhps   xmm1,xmm1
-       cvtpi2ps  xmm1,mm3
-       divps     xmm1,xmm3   ;xmm1--> | dby | dbx | dey | dex |
-       shufps    xmm1,xmm1,10110001b
-                             ;xmm1--> | dbx | dby | dex | dey |
-;1       movups    .dey23,xmm1
-       cvtps2pi  mm0,xmm1 ;mm0,xmm1          ; mm0 -> 2 delta dwords
-       movhlps   xmm1,xmm1
-       cvtps2pi  mm1,xmm1 ;mm1,xmm1
-       movq      .dey23,mm0
-       movq      .dby23,mm1
-       mov       ax,.z3
-       sub       ax,.z2
-       cwde
-       mov      dx,.x3
-       sub      dx,.x2
-       movsx    edx,dx
-       movd       mm2,[.t_x2]
-       movd       mm3,[.t_x3]
-       punpcklwd  mm2,mm4
-       punpcklwd  mm3,mm4
-       psubd      mm3,mm2
-       cvtsi2ss  xmm1,eax
-       movlhps   xmm1,xmm1
-       cvtsi2ss  xmm1,edx
-       shufps    xmm1,xmm1,00101111b
-       cvtpi2ps  xmm1,mm3
-       divps     xmm1,xmm3   ; xmm1--> | dx | dz | dty | dtx |
-       shufps    xmm1,xmm1,11100001b
-                            ; xmm1--> | dx | dz | dtx | dty |
-;       movlps    .dty23,xmm1
-;       movhps    .dz23,xmm1
-       cvtps2pi  mm0,xmm1    ; mm0 -> 2 delta dwords  | dtx | dty |
-       movhlps   xmm1,xmm1
-       cvtps2pi  mm1,xmm1    ; mm1 --> 2 delta dwords | dx  |  dz |
-       movq      .dty23,mm0
-       movq      .dz23,mm1
-else
-       mov      ax,.x3
-       sub      ax,.x2
-       cwde
-       shl      eax,ROUND
-       cdq
-       idiv     ebx
- ;     mov      .dx23,eax
-       push      eax
-       mov     ax,.z3
-       sub     ax,.z2
-       cwde
-       shl     eax,CATMULL_SHIFT
-       cdq
-       idiv    ebx
-     ; mov     .dz23,eax
-       push    eax
-       mov      ax,word[.b_x3]
-       sub      ax,word[.b_x2]
-       cwde
-       shl      eax,ROUND
-       cdq
-       idiv     ebx
- ;     mov      .dbx23,eax
-       push      eax
-       mov      ax,word[.b_y3]
-       sub      ax,word[.b_y2]
-       cwde
-       shl      eax,ROUND
-       cdq
-       idiv     ebx
- ;     mov      .dby23,eax
-       push      eax
-       mov      ax,word[.e_x3]
-       sub      ax,word[.e_x2]
-       cwde
-       shl      eax,ROUND
-       cdq
-       idiv     ebx
- ;     mov      .dex23,eax
-       push      eax
-       mov      ax,word[.e_y3]
-       sub      ax,word[.e_y2]
-       cwde
-       shl      eax,ROUND
-       cdq
-       idiv     ebx
- ;     mov      .dey23,eax
-       push      eax
-       mov      ax,word[.t_x3]
-       sub      ax,word[.t_x2]
-       cwde
-       shl      eax,ROUND
-       cdq
-       idiv     ebx
- ;     mov      .dtx23,eax
-       push      eax
-       mov      ax,word[.t_y3]
-       sub      ax,word[.t_y2]
-       cwde
-       shl      eax,ROUND
-       cdq
-       idiv     ebx
- ;     mov      .dty23,eax
-       push      eax
-end if
-      ;  sub     esp,40
-   .bt_dx23_done:
-       sub       esp,64
-       movsx    eax,.x1
-       shl      eax,ROUND
-       mov      .cx1,eax
-       mov      .cx2,eax
-  ;     push     eax
-  ;     push     eax
-       movsx    ebx,word[.b_x1]
-       shl      ebx,ROUND
-       mov      .cbx1,ebx
-       mov      .cbx2,ebx
-      ; push     ebx
-      ; push     ebx
-       movsx    ecx,word[.b_y1]
-       shl      ecx,ROUND
-       mov      .cby1,ecx
-       mov      .cby2,ecx
-      ; push     ecx
-      ; push     ecx
-       movsx    edx,word[.e_x1]
-       shl      edx,ROUND
-       mov      .cex1,edx
-       mov      .cex2,edx
-    ;   push     edx
-    ;   push     edx
-       movsx    eax,word[.e_y1]
-       shl      eax,ROUND
-       mov      .cey1,eax
-       mov      .cey2,eax
-    ;   push     eax
-    ;   push     eax
-       movsx    ebx,.z1
-       shl      ebx,CATMULL_SHIFT
-       mov      .cz1,ebx
-       mov      .cz2,ebx
-   ;    push     ebx
-   ;    push     ebx
-      ; sub      esp,16
-       movsx    ecx,word[.t_x1]
-       shl      ecx,ROUND
-       mov      .ctx1,ecx
-       mov      .ctx2,ecx
-       ;push     ecx
-       ;push     ecx
-       movsx    edx,word[.t_y1]
-       shl      edx,ROUND
-       mov      .cty1,edx
-       mov      .cty2,edx
-      ; push     edx
-      ; push     edx
-if Ext >= SSE2
-       movups  xmm0,.cby1
-       movups  xmm1,.cty1
-       movups  xmm2,.cby2
-       movups  xmm3,.cty2
-       movups  xmm4,.dby13
-       movups  xmm5,.dty13
-       movups  xmm6,.dby12
-       movups  xmm7,.dty12
-       .scby1  equ [edi]
-       .scty1  equ [edi+16]
-       .scby2  equ [edi+32]
-       .scty2  equ [edi+48]
-       .sdby13 equ [edi+64]
-       .sdty13 equ [edi+80]
-       .sdby12 equ [edi+96]
-       .sdty12 equ [edi+128]
-       push    edi
-       mov     edi,sse_repository
-       movaps  .scby1,xmm0
-       movaps  .scty1,xmm1
-       movaps  .scby2,xmm2
-       movaps  .scty2,xmm3
-       movaps  .sdby13,xmm4
-       movaps  .sdty13,xmm5
-       movaps  .sdby12,xmm6
-       movaps  .sdty12,xmm7
-       pop     edi
-end if
-       movsx    ecx,.y1
-       cmp      cx,.y2
-       jge      .loop12_done
-  .loop12:
-;if Ext >= SSE2
-;       fxsave  [sse_repository]
-;end if
-       call     .call_line
-if Ext >= SSE2
-;       fxrstor [sse_repository]
-       movups  xmm0,.cby1
-       movups  xmm1,.cty1
-       movups  xmm2,.cby2
-       movups  xmm3,.cty2
-    ;   movups  xmm4,.dby13
-    ;   movups  xmm5,.dty13
-    ;   movups  xmm6,.dby12
-    ;   movups  xmm7,.dty12
-    ;   paddd   xmm0,xmm4
-    ;   paddd   xmm1,xmm5
-    ;   paddd   xmm2,xmm6
-    ;   paddd   xmm3,xmm7
-       push    edi
-       mov     edi,sse_repository
-       paddd   xmm0,.sdby13
-       paddd   xmm1,.sdty13
-       paddd   xmm2,.sdby12
-       paddd   xmm3,.sdty12
-       pop     edi
-       movups  .cby1,xmm0
-       movups  .cty1,xmm1
-       movups  .cby2,xmm2
-       movups  .cty2,xmm3
-end if
-if (Ext = MMX) | (Ext = SSE)
-       movq     mm0,.cby2
-       movq     mm1,.cby1
-       movq     mm2,.cey2
-       movq     mm3,.cey1
-       movq     mm4,.cty1
-       movq     mm5,.cty2
-       movq     mm6,.cz1
-       movq     mm7,.cz2
-       paddd    mm0,.dby12
-       paddd    mm1,.dby13
-       paddd    mm2,.dey12
-       paddd    mm3,.dey13
-       paddd    mm4,.dty13
-       paddd    mm5,.dty12
-       paddd    mm6,.dz13
-       paddd    mm7,.dz12
-       movq     .cby2,mm0
-       movq     .cby1,mm1
-       movq     .cey1,mm3
-       movq     .cey2,mm2
-       movq     .cty1,mm4
-       movq     .cty2,mm5
-       movq     .cz1,mm6
-       movq     .cz2,mm7
-end if
-if Ext = NON
-       mov      edx,.dbx13
-       add      .cbx1,edx
-       mov      eax,.dbx12
-       add      .cbx2,eax
-       mov      ebx,.dby13
-       add      .cby1,ebx
-       mov      edx,.dby12
-       add      .cby2,edx
-       mov      eax,.dex13
-       add      .cex1,eax
-       mov      ebx,.dex12
-       add      .cex2,ebx
-       mov      edx,.dey13
-       add      .cey1,edx
-       mov      eax,.dey12
-       add      .cey2,eax
-       mov      eax,.dtx13
-       add      .ctx1,eax
-       mov      ebx,.dtx12
-       add      .ctx2,ebx
-       mov      edx,.dty13
-       add      .cty1,edx
-       mov      eax,.dty12
-       add      .cty2,eax
-       mov      eax,.dx13
-       add      .cx1,eax
-       mov      ebx,.dx12
-       add      .cx2,ebx
-       mov      ebx,.dz13
-       add      .cz1,ebx
-       mov      edx,.dz12
-       add      .cz2,edx
-end if
-       inc      ecx
-       cmp      cx,.y2
-       jl       .loop12
-    .loop12_done:
-       movsx    ecx,.y2
-       cmp      cx,.y3
-       jge      .loop23_done
-       movsx    eax,.z2
-       shl      eax,CATMULL_SHIFT
-       mov      .cz2,eax
-       movsx    ebx,.x2
-       shl      ebx,ROUND
-       mov      .cx2,ebx
-       movzx    edx,word[.b_x2]
-       shl      edx,ROUND
-       mov      .cbx2,edx
-       movzx    eax,word[.b_y2]
-       shl      eax,ROUND
-       mov      .cby2,eax
-       movzx    ebx,word[.e_x2]
-       shl      ebx,ROUND
-       mov      .cex2,ebx
-       movzx    edx,word[.e_y2]
-       shl      edx,ROUND
-       mov      .cey2,edx
-       movzx    eax,word[.t_x2]
-       shl      eax,ROUND
-       mov      .ctx2,eax
-       movzx    ebx,word[.t_y2]
-       shl      ebx,ROUND
-       mov      .cty2,ebx
-if Ext >= SSE2
-       movups  xmm2,.cby2
-       movups  xmm3,.cty2
-   ;    movups  xmm4,.dby13
-   ;    movups  xmm5,.dty13
-       movups  xmm6,.dby23
-       movups  xmm7,.dty23
-;       .scby1  equ [edi]
-;       .scty1  equ [edi+16]
-;       .scby2  equ [edi+32]
-;       .scty2  equ [edi+48]
-;       .sdby13 equ [edi+64]
-;       .sdty13 equ [edi+80]
-       .sdby23 equ [edi+160]
-       .sdty23 equ [edi+192]
-       push    edi
-       mov     edi,sse_repository
-;       movaps  .scby1,xmm0
-;       movaps  .scty1,xmm1
-       movaps  .scby2,xmm2
-       movaps  .scty2,xmm3
-;       movaps  .sdby13,xmm4
-;       movaps  .sdty13,xmm5
-       movaps  .sdby23,xmm6
-       movaps  .sdty23,xmm7
-       pop     edi
-end if
-     .loop23:
-;if Ext >= SSE2
-;       fxsave  [sse_repository]
-;end if
-       call     .call_line
-if Ext >= SSE2
-       movups  xmm0,.cby1
-       movups  xmm1,.cty1
-       movups  xmm2,.cby2
-       movups  xmm3,.cty2
-       push    edi
-       mov     edi,sse_repository
-       paddd   xmm0,.sdby13
-       paddd   xmm1,.sdty13
-       paddd   xmm2,.sdby23
-       paddd   xmm3,.sdty23
-       pop     edi
-       movups  .cby1,xmm0
-       movups  .cty1,xmm1
-       movups  .cby2,xmm2
-       movups  .cty2,xmm3
-;       fxrstor [sse_repository]
-;       movups  xmm0,.cby1
-;       movups  xmm1,.cty1
-;       movups  xmm2,.cby2
-;       movups  xmm3,.cty2
-;       movups  xmm4,.dby13
-;       movups  xmm5,.dty13
-;       movups  xmm6,.dby23
-;       movups  xmm7,.dty23
-;       paddd   xmm0,xmm4
-;       paddd   xmm1,xmm5
-;       paddd   xmm2,xmm6
- ;      paddd   xmm3,xmm7
- ;      movups  .cby1,xmm0
- ;      movups  .cty1,xmm1
- ;      movups  .cby2,xmm2
- ;      movups  .cty2,xmm3
-;
-end if
-if (Ext = MMX) | (Ext = SSE)
-       movq     mm0,.cby2
-       movq     mm1,.cby1
-       movq     mm2,.cey2
-       movq     mm3,.cey1
-       movq     mm4,.cty1
-       movq     mm5,.cty2
-       movq     mm6,.cz1
-       movq     mm7,.cz2
-       paddd    mm0,.dby23
-       paddd    mm1,.dby13
-       paddd    mm2,.dey23
-       paddd    mm3,.dey13
-       paddd    mm4,.dty13
-       paddd    mm5,.dty23
-       paddd    mm6,.dz13
-       paddd    mm7,.dz23
-       movq     .cby2,mm0
-       movq     .cby1,mm1
-       movq     .cey2,mm2
-       movq     .cey1,mm3
-       movq     .cty1,mm4
-       movq     .cty2,mm5
-       movq     .cz1,mm6
-       movq     .cz2,mm7
-end if
-If Ext = NON
-       mov      edx,.dbx13
-       add      .cbx1,edx
-       mov      eax,.dbx23
-       add      .cbx2,eax
-       mov      ebx,.dby13
-       add      .cby1,ebx
-       mov      edx,.dby23
-       add      .cby2,edx
-       mov      eax,.dex13
-       add      .cex1,eax
-       mov      ebx,.dex23
-       add      .cex2,ebx
-       mov      edx,.dey13
-       add      .cey1,edx
-       mov      eax,.dey23
-       add      .cey2,eax
-       mov      eax,.dx13
-       add      .cx1,eax
-       mov      ebx,.dx23
-       add      .cx2,ebx
-       mov      ebx,.dz13
-       add      .cz1,ebx
-       mov      edx,.dz23
-       add      .cz2,edx
-       mov      eax,.dtx13
-       add      .ctx1,eax
-       mov      ebx,.dtx23
-       add      .ctx2,ebx
-       mov      edx,.dty13
-       add      .cty1,edx
-       mov      eax,.dty23
-       add      .cty2,eax
-end if
-       inc      ecx
-       cmp      cx,.y3
-       jl       .loop23
-    .loop23_done:
-       mov      esp,ebp
-ret   50
-.call_line:
-       pushad
-       ; xmm0= cby1,cbx1,cz1,cx1
-       ; xmm1= cty1,ctx1,cey1,cex1
-if Ext >= SSE2
-       sub      esp,8
-       shufps   xmm1,xmm1,10110001b
-       shufps   xmm3,xmm3,10110001b
-       movlps   [esp],xmm1
-else
-       push     dword .cty1
-       push     .ctx1
-end if
-       push     dword .cz1
-if Ext>=SSE2
-       sub      esp,8
-       movlps   [esp],xmm3
-else
-       push     dword .cty2
-       push     .ctx2
-end if
-       push     dword .cz2
-if Ext>=SSE2
-       sub      esp,32
-       movhps   [esp+24],xmm3
-       shufps   xmm2,xmm2,10110001b
-       movlps   [esp+16],xmm2
-       movhps   [esp+8],xmm1
-       shufps   xmm0,xmm0,10110001b
-       movlps   [esp],xmm0 ;================================
-else
-       push     dword .cey2
-       push     .cex2
-       push     dword .cby2
-       push     .cbx2
-       push     dword .cey1
-       push     .cex1
-       push     dword .cby1
-       push     .cbx1
-end if
-       push     .tex_ptr
-       push     .z_buff
-       push     .t_emap
-       push     .t_bmap
-       push     ecx
-       mov      eax,.cx1
-       sar      eax,ROUND
-       mov      ebx,.cx2
-       sar      ebx,ROUND
-       call     bump_tex_line_z
-       popad
-;end if
-ret
-bump_tex_line_z:
-;--------------in: eax - x1
-;--------------    ebx - x2
-;--------------    edi - pointer to screen buffer
-;stack - another parameters :
-.y      equ dword [ebp+4]
-.bmap   equ dword [ebp+8]        ; bump map pointer
-.emap   equ dword [ebp+12]       ; env map pointer
-.z_buff equ dword [ebp+16]       ; z buffer
-.tex_map equ dword [ebp+20]      ; texture pointer
-.bx1    equ  [ebp+24]   ;   ---
-.by1    equ  [ebp+28]  ;       |
-.ex1    equ  [ebp+32]  ;       |
-.ey1    equ  [ebp+36]  ;       |
-.bx2    equ  [ebp+40]  ;       |
-.by2    equ  [ebp+44]  ;       |>   b. map and e. map coords
-.ex2    equ  [ebp+48]  ;       |>   shifted shl ROUND
-.ey2    equ  [ebp+52]  ;   ---
-.z2     equ  [ebp+56]
-.tx2    equ  [ebp+60]
-.ty2    equ  [ebp+64]
-.z1     equ  [ebp+68]
-.tx1    equ  [ebp+72]
-.ty1    equ  [ebp+76]
-.x1     equ [ebp-4]
-.x2     equ [ebp-8]
-.dbx    equ [ebp-12]
-.dby    equ [ebp-16]
-.dex    equ [ebp-20]
-.dey    equ [ebp-24]
-.dz     equ [ebp-28]
-.dtx    equ [ebp-32]
-.dty    equ [ebp-36]
-.cbx    equ [ebp-40]
-.cby    equ [ebp-44]
-.cex    equ [ebp-48]
-.cey    equ [ebp-52]
-.cz     equ [ebp-56]
-.czbuff equ [ebp-60]
-.ctx    equ [ebp-64]
-.cty    equ [ebp-68]
-.c_scr  equ [ebp-72]
-.temp1  equ        ebp-80
-.temp2  equ        ebp-88
-.temp3  equ        ebp-76
-.temp4  equ        ebp-84
-.temp5  equ        ebp-92
-        mov     ebp,esp
-        mov     ecx,.y
-        or      ecx,ecx
-        jl      .bl_end
-        movzx   edx,word[size_y_var]
-        cmp     ecx,edx  ;SIZE_Y
-        jge     .bl_end
-        cmp     eax,ebx
-        jl      .bl_ok
-        je      .bl_end
-if Ext=NON
-        mov     edx,.bx1
-        xchg    edx,.bx2
-        mov     .bx1,edx
-        mov     edx,.by1
-        xchg    edx,.by2
-        mov     .by1,edx
-        mov     edx,.ex1
-        xchg    edx,.ex2
-        mov     .ex1,edx
-        mov     edx,.ey1
-        xchg    edx,.ey2
-        mov     .ey1,edx
-        mov     edx,.tx1
-        xchg    edx,.tx2
-        mov     .tx1,edx
-        mov     edx,.ty1
-        xchg    edx,.ty2
-        mov     .ty1,edx
-end if
-if Ext = MMX
-        movq    mm0,.bx1
-        movq    mm1,.bx2
-        movq    mm2,.ex1
-        movq    mm3,.ex2
-        movq    mm4,.tx1
-        movq    mm5,.tx2
-        movq    .bx2,mm0
-        movq    .bx1,mm1
-        movq    .ex1,mm3
-        movq    .ex2,mm2
-        movq    .tx1,mm5
-        movq    .tx2,mm4
-end if
-if Ext>=SSE
-        movups xmm0,.bx1
-        movups xmm1,.bx2
-        movups .bx1,xmm1
-        movups .bx2,xmm0
-        movq    mm0,.tx1
-        movq    mm1,.tx2
-        movq    .tx1,mm1
-        movq    .tx2,mm0
-end if
-;if Ext>=SSE2
-;        movaps  xmm4,xmm0
-;        movaps  xmm0,xmm2
-;        movaps  xmm2,xmm4
-;        movaps  xmm5,xmm1
-;        movaps  xmm1,xmm3
-;        movaps  xmm3,xmm5
-;else
-        xchg    eax,ebx
-        mov     edx,.z1
-        xchg    edx,.z2
-        mov     .z1,edx
-;end if
-  .bl_ok:
-;if Ext >= SSE2
-;        shufps  xmm0,xmm0,11100001b
-;        shufps  xmm2,xmm2,11100001b
-;        movlps  .bx1,xmm0
-;        movlps  .bx2,xmm2
-;        shufps  xmm0,xmm0,00011011b
-;        shufps  xmm2,xmm2,00011011b
-;        movd    eax,xmm0
-;        movd    ebx,xmm2
-;        shufps  xmm0,xmm0,11000110b
-;        shufps  xmm2,xmm2,11000110b
-;        movd    .z1,xmm0
-;        movd    .z2,xmm2
-;        shufps  xmm1,xmm1,10110001b
-;        shufps  xmm3,xmm3,10110001b
-;        movlps  .ex1,xmm1
-;        movlps  .ex2,xmm2
-;        movhps  .tx1,xmm1
-;        movhps  .tx2,xmm2
-;        xchg    eax,ebx
-;        mov     edx,.z1
-;        xchg    edx,.z2
-;        mov     .z1,edx
-;end if
-        push    eax
-        push    ebx           ;store x1, x2
-        movzx   ebx,word[size_x_var]
-    ;    mov     eax,.x1
-        cmp     dword .x1,ebx  ;dword .x1,SIZE_X
-        jge     .bl_end
-        cmp     dword .x2,0
-        jle     .bl_end
-        mov     ebx,.x2
-        sub     ebx,.x1
-if Ext>=SSE
-       sub       esp,28
-       cvtsi2ss  xmm3,ebx            ;rcps
-       shufps    xmm3,xmm3,0
-; float using SSE variant  ::-->
-;       movups    xmm0,.bx1  ; new
-;       movups    xmm1,.bx2  ; new
-       cvtpi2ps  xmm0,.bx1 ;mm0    ; variant fixed point
-       movlhps   xmm0,xmm0
-       cvtpi2ps  xmm0,.ex1 ;mm2
-       cvtpi2ps  xmm1,.bx2 ;mm1
-       movlhps   xmm1,xmm1
-       cvtpi2ps  xmm1,.ex2 ;mm3
-       subps     xmm1,xmm0
-       divps     xmm1,xmm3
-       shufps    xmm1,xmm1,10110001b
-;       movups    .dey,xmm1  ; new
-       cvtps2pi  mm0,xmm1          ; mm0 -> 2 delta dwords
-       movhlps   xmm1,xmm1
-       cvtps2pi  mm1,xmm1
-       movq      .dey,mm0
-       movq      .dby,mm1
-       movd      mm2,.z1
-       movd      mm3,.z2
-       cvtpi2ps  xmm0,.tx1 ;mm0
-       movlhps   xmm0,xmm0
-       cvtpi2ps  xmm0,mm2
-       cvtpi2ps  xmm1,.tx2 ;mm1
-       movlhps   xmm1,xmm1
-       cvtpi2ps  xmm1,mm3
-;       movups    xmm0,,z1  ; new
-;       movups    xmm1,.z2  ; new
-       subps     xmm1,xmm0
-       divps     xmm1,xmm3
-;       movups    .dz,xmm1  ;new
-       shufps    xmm1,xmm1,10110100b
-       cvtps2pi  mm0,xmm1          ; mm0 -> 2 delta dwords
-       movhlps   xmm1,xmm1
-       cvtps2pi  mm1,xmm1
-       movd      .dz,mm0
-       movq      .dty,mm1
-else
-        mov     eax,.bx2       ; calc .dbx
-        sub     eax,.bx1
-        cdq
-        idiv    ebx
-        push    eax
-        mov     eax,.by2       ; calc .dby
-        sub     eax,.by1
-        cdq
-        idiv    ebx
-        push    eax
-        mov     eax,.ex2       ; calc .dex
-        sub     eax,.ex1
-        cdq
-        idiv    ebx
-        push    eax
-        mov     eax,.ey2       ; calc .dey
-        sub     eax,.ey1
-        cdq
-        idiv    ebx
-        push    eax
-        mov     eax,.z2        ; calc .dz
-        sub     eax,.z1
-        cdq
-        idiv    ebx
-        push    eax
-        mov     eax,.tx2       ; calc .dtx
-        sub     eax,.tx1
-        cdq
-        idiv    ebx
-        push    eax
-        mov     eax,.ty2       ; calc .dty
-        sub     eax,.ty1
-        cdq
-        idiv    ebx
-        push    eax
-end if
-        cmp     dword .x1,0         ; set correctly begin variable
-        jge     @f            ; CLIPPING ON FUNCTION
-                              ; cutting triangle exceedes screen
-        mov     ebx,.x1
-        neg     ebx
-;if Ext >= SSE
-;        cvtsi2ss xmm0,ebx
-;        shufps   xmm0,xmm0,0
-;        movups   xmm1,.dey
-;        mulps    xmm1,xmm0
-;        shufps   xmm1,xmm1,00011011b
-;        movups   xmm2,.bx1
-;        addps    xmm2,xmm1
-;        movups   .bx1,xmm2
-        mov     eax,.dz
-        imul    ebx           ; eax = .dz * abs(.x1)
-        add     .z1,eax
-        mov     dword .x1,0
-        mov     eax,.dbx
-        imul    ebx
-        add    .bx1,eax
-        mov     eax,.dby
-        imul    ebx
-        add     .by1,eax
-        mov     eax,.dex
-        imul    ebx
-        add     .ex1,eax
-        mov     eax,.dey
-        imul    ebx
-        add     .ey1,eax
-        mov     eax,.dtx
-        imul    ebx
-        add     .tx1,eax
-        mov     eax,.dty
-        imul    ebx
-        add     .ty1,eax
-      @@:
-   ;     mov     ebx,.x2
-        movzx   eax,word[size_x_var]
-       ; cmp     dword .x2,SIZE_X
-        cmp     dword .x2,eax  ; eax,ebx
-        jl      @f
-        mov     dword .x2,eax  ;SIZE_X
-      @@:
-        movzx   eax,word[size_x_var]  ;SIZE_X       ;calc memory begin in buffers
-        mul     .y
-        add     eax,.x1
-        lea     esi,[4*eax]
-        add     esi,.z_buff       ; z-buffer filled with dd variables
-        lea     eax,[eax*3]
-        add     edi,eax
-        mov     ecx,.x2
-        sub     ecx,.x1
-        ; init current variables
-        push    dword .bx1   ; current b, e and t shifted shl ROUND   .cbx
-        push    dword .by1                                         ;  .cby
-        push    dword .ex1                                         ;  .cex
-        push    dword .ey1                                         ;  .cey
-        push    dword .z1    ; current z shl CATMULL_SHIFT         ; .cz
-        push    esi                                          ; .czbuff
-        push    dword .tx1      ;         .ctx
-        push    dword .ty1      ;         .cty
-        push    edi       ;         .c_scr
-if Ext = SSE2
-        mov    eax,TEXTURE_SIZE
-        movd   xmm1,eax
-        shufps xmm1,xmm1,0
-        push   dword  TEX_X
-        push   dword  -TEX_X
-        push   dword  1
-        push   dword  -1
-        movups xmm2,[esp]
-        movd   xmm3,.bmap
-        shufps xmm3,xmm3,0
-end if
-if Ext>=MMX
-        movq    mm7,.cty
-        movq    mm6,.cby
-        movq    mm5,.cey
-;        movq    mm4,.dtyq
-;        movq    mm3,.dbyq
-end if
-     .draw:
-    ; if TEX = SHIFTING   ;bump drawing only in shifting mode
-        mov     esi,.czbuff      ; .czbuff current address in buffer
-        mov     ebx,.cz          ; .cz - cur z position
-        cmp     ebx,dword[esi]
-        jge     .skip
-if Ext=NON
-        mov     eax,.cby
-        shr     eax,ROUND
-        mov     esi,.cbx
-        shr     esi,ROUND
-else
-        movq    mm1,mm6
-        psrld   mm1,ROUND
-        movd    eax,mm1
-        psrlq   mm1,32
-        movd    esi,mm1
-end if
-        shl     eax,TEX_SHIFT
-        add     esi,eax         ;-  ; esi - current bump map index
-if Ext = SSE2
-        movd    xmm0,esi
-        shufps  xmm0,xmm0,0
-        paddd   xmm0,xmm2
-        pand    xmm0,xmm1
-        paddd   xmm0,xmm3
-        movd    ebx,xmm0
-        movzx   eax,byte[ebx]
-;
-;        shufps  xmm0,xmm0,11100001b
-        psrldq  xmm0,4
-        movd    ebx,xmm0
-        movzx   ebx,byte[ebx]
-        sub     eax,ebx
-;
-;        shufps  xmm0,xmm0,11111110b
-        psrldq  xmm0,4
-        movd    ebx,xmm0
-        movzx   edx, byte [ebx]
-;
-;        shufps  xmm0,xmm0,11111111b
-        psrldq  xmm0,4
-        movd    ebx,xmm0
-        movzx   ebx, byte [ebx]
-        sub     edx,ebx
-;
-else
-;        mov     ebx,esi
-;        dec     ebx
-        lea     ebx,[esi-1]
-        and     ebx,TEXTURE_SIZE
-        add     ebx,.bmap
-        movzx   eax,byte [ebx]
-;        mov     ebx,esi
-;        inc     ebx
-        lea     ebx,[esi+1]
-        and     ebx,TEXTURE_SIZE
-        add     ebx,.bmap
-        movzx   ebx,byte [ebx]
-        sub     eax,ebx
-;        mov     ebx,esi
-;        sub     ebx,TEX_X
-        lea     ebx,[esi-TEX_X]
-        and     ebx,TEXTURE_SIZE
-        add     ebx,.bmap
-        movzx   edx,byte [ebx]
-;        mov     ebx,esi
-;        add     ebx,TEX_X
-        lea     ebx,[esi+TEX_X]
-        and     ebx,TEXTURE_SIZE
-        add     ebx,.bmap
-        movzx   ebx,byte [ebx]
-        sub     edx,ebx
-end if
-     ;  eax - horizontal sub    modificated x coord
-     ;  edx - vertical   sub    modificated y coord
-if Ext=NON
-        mov     ebx,.cex       ;.cex - current env map X
-        shr     ebx,ROUND
-        add     eax,ebx
-        mov     ebx,.cey       ;.cey - current  env map y
-        shr     ebx,ROUND
-        add     edx,ebx
-else
-        movq    mm1,mm5        ; mm5 - copy of cur env coords
-        psrld   mm1,ROUND
-        movd    ebx,mm1
-        psrlq   mm1,32
-        add     eax,ebx
-        movd    ebx,mm1
-        add     edx,ebx
-;        movq    qword[.temp1],mm3
-;        add     eax,dword [.temp1]
-;        add     edx,dword [.temp1+4]
-end if
-        or      eax,eax
-        jl      .black
-        cmp     eax,TEX_X
-        jg      .black
-        or      edx,edx
-        jl      .black
-        cmp     edx,TEX_Y
-        jg      .black
-        shl     edx,TEX_SHIFT   ; zaburzenie w emapie = zaburzenie w teksturze
-        add     edx,eax         ; proponuje nie stawiac czarnego pixela tylko
-        lea     esi,[edx*3]     ; niezaburzony.
-        add     esi,.emap       ;
-        lodsd
-if Ext=NON
-        mov     edx,.cty
-        shr     edx,ROUND  ; sar
-        mov     edi,.ctx
-        shr     edi,ROUND  ; sar
-else
-        movq    mm1,mm7
-        psrld   mm1,ROUND
-        movd    edx,mm1
-        psrlq   mm1,32
-        movd    edi,mm1
-end if
-        shl     edx,TEX_SHIFT
-        add     edi,edx
-        and     edi,TEXTURE_SIZE
-        lea     esi,[edi*3]
-        add     esi,.tex_map
-if Ext=NON
-        mov     edx,eax
-        lodsd
-        push    ax
-        mul     dl
-        mov     dl,ah
-        pop     ax
-        shr     ax,8
-        mul     dh
-        mov     al,dl
-        mov     edi,.c_scr
-        stosw
-        shr     edx,16
-        shr     eax,16
-        mul     dl
-        shr     ax,8
-        stosb
-else
-        movd       mm0,eax
-        pxor       mm1,mm1
-        punpcklbw  mm0,mm1
-        movd       mm2,[esi]
-        punpcklbw  mm2,mm1
-        pmullw     mm0,mm2
-        psrlw      mm0,8
-        packuswb   mm0,mm1
-        mov        edi,.c_scr
-        movd       [edi],mm0
-end if
-        jmp     .actual_zbuff   ; actualize z buffer
-     @@:
-     .black:
-        xor     eax,eax
-        mov     edi,.c_scr
-        stosd
-     .actual_zbuff:
-        mov     eax,.cz
-        mov     edi,.czbuff
-        stosd
-      .skip:
-        add     dword .czbuff,4
-        add     dword .c_scr,3
-if Ext=NON
-        mov     eax,.dbx
-        add     .cbx,eax
-        mov     ebx,.dby
-        add     .cby,ebx
-        mov     edx,.dex
-        add     .cex,edx
-        mov     eax,.dey
-        add     .cey,eax
-        mov     ebx,.dtx
-        add     .ctx,ebx
-        mov     edx,.dty
-        add     .cty,edx
-else
-        paddd   mm7,.dty
-        paddd   mm6,.dby
-        paddd   mm5,.dey
-end if
-        mov     eax,.dz
-        add     .cz,eax
-        dec     ecx
-        jnz     .draw
-  .bl_end:
-        mov     esp,ebp
-ret 76
-;Ext = MMX
-;     else
-;        movq    mm5, qword[.temp1]  ;-
-;        paddd   mm5, qword[.temp5]  ; .temp5 == low dword = TEX_X, high dword = -TEX_X
-;        pand    mm5, qword[.temp3]  ; .temp3 == low = high dword = TEX_SIZE
-;        paddd   mm5, qword[.temp4]  ; .temp4 == low = high dword = offset .bmap
-;        movd    ebx,mm5
-;        psrlq   mm5,32
-;     end if
+;CATMULL_SHIFT equ 8
+;TEXTURE_SIZE EQU (TEX_X * TEX_Y)-1
+;ROUND equ 8
+;Ext = NON
+;MMX = 1
+;NON = 0
+;------- Big thanks to Majuma (www.majuma.xt.pl) for absolutely great---
+;------- DOS 13h mode demos --------------------------------------------
+;------- Procedure draws bump triangle with texture, I use -------------
+;--------Catmull Z-buffer algorithm- (Z coordinate interpolation)-------
+;--------I calc texture pixel by this way: col1*col2/256 ---------------
+bump_tex_triangle_z:
+;------------------in - eax - x1 shl 16 + y1 -----------
+;---------------------- ebx - x2 shl 16 + y2 -----------
+;---------------------- ecx - x3 shl 16 + y3 -----------
+;---------------------- edx - pointer to bump map-------
+;---------------------- esi - pointer to env map--------
+;---------------------- edi - pointer to screen buffer--
+;---------------------- stack : bump coordinates--------
+;----------------------         environment coordinates-
+;----------------------         Z position coordinates--
+;----------------------         pointer to Z buffer-----
+;----------------------         pointer to texture------
+;----------------------         texture coordinates-----
+;-- Z-buffer - filled with coordinates as dword --------
+;-- (Z coor. as word) shl CATMULL_SHIFT ----------------
+.b_x1   equ ebp+4   ; procedure don't save registers !!!
+.b_y1   equ ebp+6   ; each coordinate as word
+.b_x2   equ ebp+8
+.b_y2   equ ebp+10       ; b - bump map coords
+.b_x3   equ ebp+12       ; e - env map coords
+.b_y3   equ ebp+14
+.e_x1   equ ebp+16
+.e_y1   equ ebp+18
+.e_x2   equ ebp+20
+.e_y2   equ ebp+22
+.e_x3   equ ebp+24
+.e_y3   equ ebp+26
+.z1     equ word[ebp+28]
+.z2     equ word[ebp+30]
+.z3     equ word[ebp+32]
+.z_buff equ dword[ebp+34]       ; pointer to Z-buffer
+.tex_ptr equ dword[ebp+38]      ; ptr to texture
+.t_x1   equ ebp+42              ; texture coords
+.t_y1   equ ebp+44
+.t_x2   equ ebp+46
+.t_y2   equ ebp+48
+.t_x3   equ ebp+50
+.t_y3   equ ebp+52
+.t_bmap equ dword[ebp-4]        ; pointer to bump map
+.t_emap equ dword[ebp-8]        ; pointer to env map
+.x1     equ word[ebp-10]
+.y1     equ word[ebp-12]
+.x2     equ word[ebp-14]
+.y2     equ word[ebp-16]
+.x3     equ word[ebp-18]
+.y3     equ word[ebp-20]
+if 0 ;Ext <= SSE2
+.dx12   equ dword[edi-4]
+.dz12   equ      [edi-8]
+.dbx12  equ dword[edi-12]
+.dby12  equ      [edi-16]
+.dex12  equ dword[edi-20]
+.dey12  equ      [edi-24]
+.dtx12  equ dword[edi-28]
+.dty12  equ      [edi-32]
+.dx13  equ dword[ebp-52-4*1]
+.dz13  equ      [ebp-52-4*2]
+.dbx13 equ dword[ebp-52-4*3]
+.dby13 equ      [ebp-52-4*4]
+.dex13 equ dword[ebp-52-4*5]
+.dey13 equ      [ebp-52-4*6]
+.dtx13 equ dword[ebp-52-4*7]
+.dty13 equ      [ebp-52-4*8]
+.dx23  equ dword[ebp-(52+4*9)]
+.dz23  equ      [ebp-(52+4*10)]
+.dbx23 equ dword[ebp-(52+4*11)]
+.dby23 equ      [ebp-(52+4*12)]
+.dex23 equ dword[ebp-(52+4*13)]
+.dey23 equ      [ebp-(52+4*14)]
+.dtx23 equ dword[ebp-(52+4*15)]
+.dty23 equ      [ebp-(52+4*16)]
+else
+.dx12   equ dword[ebp-24]
+.dz12   equ      [ebp-28]
+.dbx12  equ dword[ebp-32]
+.dby12  equ      [ebp-36]
+.dex12  equ dword[ebp-40]
+.dey12  equ      [ebp-44]
+.dtx12  equ dword[ebp-48]
+.dty12  equ      [ebp-52]
+.dx13  equ dword[ebp-52-4*1]
+.dz13  equ      [ebp-52-4*2]
+.dbx13 equ dword[ebp-52-4*3]
+.dby13 equ      [ebp-52-4*4]
+.dex13 equ dword[ebp-52-4*5]
+.dey13 equ      [ebp-52-4*6]
+.dtx13 equ dword[ebp-52-4*7]
+.dty13 equ      [ebp-52-4*8]
+.dx23  equ dword[ebp-(52+4*9)]
+.dz23  equ      [ebp-(52+4*10)]
+.dbx23 equ dword[ebp-(52+4*11)]
+.dby23 equ      [ebp-(52+4*12)]
+.dex23 equ dword[ebp-(52+4*13)]
+.dey23 equ      [ebp-(52+4*14)]
+.dtx23 equ dword[ebp-(52+4*15)]
+.dty23 equ      [ebp-(52+4*16)]
+end if
+if Ext < SSE
+.cx1   equ dword[ebp-(52+4*17)]         ; current variables
+.cz1   equ      [ebp-(52+4*18)]
+.cx2   equ dword[ebp-(52+4*19)]
+.cz2   equ      [ebp-(52+4*20)]
+.cbx1  equ dword[ebp-(52+4*21)]
+.cby1  equ      [ebp-(52+4*22)]
+.cbx2  equ dword[ebp-(52+4*23)]
+.cby2  equ      [ebp-(52+4*24)]
+.cex1  equ dword[ebp-(52+4*25)]
+.cey1  equ      [ebp-(52+4*26)]
+.cex2  equ dword[ebp-(52+4*27)]
+.cey2  equ      [ebp-(52+4*28)]
+.ctx1  equ dword[ebp-(52+4*29)]
+.cty1  equ      [ebp-(52+4*30)]
+.ctx2  equ dword[ebp-(52+4*31)]
+.cty2  equ      [ebp-(52+4*32)]
+else
+.cx1   equ dword[ebp-(52+4*17)]         ; current variables
+.cz1   equ      [ebp-(52+4*18)]
+.cbx1  equ dword[ebp-(52+4*19)]
+.cby1  equ      [ebp-(52+4*20)]
+.cex1  equ dword[ebp-(52+4*21)]
+.cey1  equ      [ebp-(52+4*22)]
+.ctx1  equ dword[ebp-(52+4*23)]
+.cty1  equ      [ebp-(52+4*24)]
+.cx2   equ dword[ebp-(52+4*25)]
+.cz2   equ      [ebp-(52+4*26)]
+.cbx2  equ dword[ebp-(52+4*27)]
+.cby2  equ      [ebp-(52+4*28)]
+.cex2  equ dword[ebp-(52+4*29)]
+.cey2  equ      [ebp-(52+4*30)]
+.ctx2  equ dword[ebp-(52+4*31)]
+.cty2  equ      [ebp-(52+4*32)]
+end if
+       cld
+       mov     ebp,esp
+       push    edx        ; store bump map
+       push    esi        ; store e. map
+     ; sub     esp,120
+ .sort3:                  ; sort triangle coordinates...
+       cmp     ax,bx
+       jle     .sort1
+       xchg    eax,ebx
+       mov     edx,dword[.b_x1]
+       xchg    edx,dword[.b_x2]
+       mov     dword[.b_x1],edx
+       mov     edx,dword[.e_x1]
+       xchg    edx,dword[.e_x2]
+       mov     dword[.e_x1],edx
+       mov     edx,dword[.t_x1]
+       xchg    edx,dword[.t_x2]
+       mov     dword[.t_x1],edx
+       mov     dx,.z1
+       xchg    dx,.z2
+       mov     .z1,dx
+ .sort1:
+       cmp      bx,cx
+       jle      .sort2
+       xchg     ebx,ecx
+       mov      edx,dword[.b_x2]
+       xchg     edx,dword[.b_x3]
+       mov      dword[.b_x2],edx
+       mov      edx,dword[.e_x2]
+       xchg     edx,dword[.e_x3]
+       mov      dword[.e_x2],edx
+       mov      edx,dword[.t_x2]
+       xchg     edx,dword[.t_x3]
+       mov      dword[.t_x2],edx
+       mov     dx,.z2
+       xchg    dx,.z3
+       mov     .z2,dx
+       jmp      .sort3
+ .sort2:
+       push     eax     ; store triangle coords in variables
+       push     ebx
+       push     ecx
+         mov      edx,80008000h  ; eax,ebx,ecx are ANDd together into edx which means that
+         and      edx,ebx        ; if *all* of them are negative a sign flag is raised
+         and      edx,ecx
+         and      edx,eax
+         test     edx,80008000h  ; Check both X&Y at once
+         jne      .loop23_done
+    ;   mov     edx,eax         ; eax,ebx,ecx are ORd together into edx which means that
+    ;   or      edx,ebx         ; if any *one* of them is negative a sign flag is raised
+    ;   or      edx,ecx
+    ;   test    edx,80000000h   ; Check only X
+    ;   jne     .loop23_done
+    ;   cmp     .x1,SIZE_X    ; {
+    ;   jg      .loop23_done
+    ;   cmp     .x2,SIZE_X     ; This can be optimized with effort
+    ;   jg      .loop23_done
+    ;   cmp     .x3,SIZE_X
+    ;   jg      .loop23_done    ; {
+       mov      bx,.y2       ; calc delta 12
+       sub      bx,.y1
+       jnz      .bt_dx12_make
+if 0 ;Ext >= SSE2
+       pxor     xmm0,xmm0
+       movups   .dty12,xmm0
+       movups   .dey12,xmm0
+       sub      esp,16
+else
+       mov      ecx,8
+       xor      edx,edx
+     @@:
+       push     edx   ;dword 0
+       loop     @b
+end if
+       jmp      .bt_dx12_done
+ .bt_dx12_make:
+       movsx    ebx,bx
+if Ext>=SSE
+       sub       esp,32
+   ;    mov       eax,256
+       cvtsi2ss  xmm4,[i255d]
+       cvtsi2ss  xmm3,ebx ;rcps
+if 0 ;Ext >= SSE2
+       mov       edi,ebp
+       sub       edi,512
+       or        edi,0x0000000f
+end if
+       divss     xmm3,xmm4
+       shufps    xmm3,xmm3,0
+       movd      mm0,[.b_x1]
+       movd      mm1,[.b_x2]
+       movd      mm2,[.e_x1]
+       movd      mm3,[.e_x2]
+       pxor       mm4,mm4
+       punpcklwd  mm0,mm4
+       punpcklwd  mm1,mm4
+       punpcklwd  mm2,mm4
+       punpcklwd  mm3,mm4
+       psubd      mm1,mm0
+       psubd      mm3,mm2
+       cvtpi2ps  xmm1,mm1
+       movlhps   xmm1,xmm1
+       cvtpi2ps  xmm1,mm3
+       divps     xmm1,xmm3   ;xmm1--> | dby | dbx | dey | dex |
+       shufps    xmm1,xmm1,10110001b
+                             ;xmm1--> | dbx | dby | dex | dey |
+;1       movups    .dey12,xmm1
+       cvtps2pi  mm0,xmm1 ;mm0,xmm1          ; mm0 -> 2 delta dwords
+       movhlps   xmm1,xmm1
+       cvtps2pi  mm1,xmm1 ;mm1,xmm1
+       movq      .dey12,mm0
+       movq      .dby12,mm1
+;-------------
+  ;    pxor      mm0,mm0
+  ;    pxor      mm1,mm1
+   ;/   pinsrw    mm0,.z1,1
+   ;/   pinsrw    mm0,.x1,0
+   ;/   pinsrw    mm1,.z2,1
+   ;/   pinsrw    mm1,.x2,0
+       mov       ax,.z2
+       sub       ax,.z1
+       cwde
+       mov      dx,.x2
+       sub      dx,.x1
+       movsx    edx,dx
+   ;/    movd      mm1,eax
+   ;/    punpcklwd  mm0,mm4
+   ;/    punpcklwd  mm1,mm4
+  ;     cvtpi2ps   xmm1,mm1
+  ;     cvtpi2ps   xmm2,mm0
+  ;     subps      xmm1,xmm2
+   ;/   psubd      mm1,mm0
+       movd       mm2,[.t_x1]
+       movd       mm3,[.t_x2]
+       punpcklwd  mm2,mm4
+       punpcklwd  mm3,mm4
+       psubd      mm3,mm2
+   ;/  cvtpi2ps  xmm1,mm1
+       cvtsi2ss  xmm1,eax
+       movlhps   xmm1,xmm1
+       cvtsi2ss  xmm1,edx
+   ;    movss     xmm1,xmm4
+       shufps    xmm1,xmm1,00101111b
+       cvtpi2ps  xmm1,mm3
+       divps     xmm1,xmm3   ; xmm1--> | dx | dz | dty | dtx |
+       shufps    xmm1,xmm1,11100001b
+                             ; xmm1--> | dx | dz | dtx | dty |
+;1       movlps    .dty12,xmm1
+;1       movhps    .dz12,xmm1
+       cvtps2pi  mm0,xmm1    ; mm0 -> 2 delta dwords  | dtx | dty |
+       movhlps   xmm1,xmm1
+       cvtps2pi  mm1,xmm1
+       movq      .dty12,mm0
+       movq      .dz12,mm1
+;----
+;       mov       ax,.z2
+;       sub       ax,.z1
+;       cwde
+;       mov       bx,.x2
+;       sub       bx,.x1
+;       movsx     ebx,bx
+;       movd      mm1,eax
+;       psllq     mm1,32
+;       movd      mm1,ebx
+;;       push      ebx
+;;       push      eax
+;;       movq      mm1,[esp]
+;;       add       esp,8
+;;;       mov       ax,.z1
+;;;       mov       bx,.z2
+;;;       shl       eax,16
+;;;       shl       ebx,16
+;;;       mov       ax,.x1
+;;;       mov       bx,.x2
+;       movd       mm2,[.t_x1]
+;       movd       mm3,[.t_x2]
+;;       movd      mm0,eax
+;;       movd      mm1,ebx
+;       pxor       mm4,mm4
+;;       punpcklwd  mm0,mm4
+;;       punpcklwd  mm1,mm4
+;       punpcklwd  mm2,mm4
+;       punpcklwd  mm3,mm4
+;;       psubd    mm1,mm0
+;       psubd      mm3,mm2
+;       cvtpi2ps  xmm1,mm1
+;       movlhps   xmm1,xmm1
+;       cvtpi2ps  xmm1,mm3
+;       divps     xmm1,xmm3   ; xmm1--> | dz | dx | dty | dtx |
+;       shufps    xmm1,xmm1,10110001b
+                             ; xmm1--> | dx | dz | dtx | dty |
+;       cvtps2pi  mm0,xmm1    ; mm0 -> 2 delta dwords  | dtx | dty |
+;       movhlps   xmm1,xmm1
+;       cvtps2pi  mm1,xmm1    ; mm1 --> 2 delta dwords | dx | dz |
+;       movq      .dty12,mm0
+;       movq      .dz12,mm1
+else
+       mov      ax,.x2
+       sub      ax,.x1
+       cwde
+       shl      eax,ROUND
+       cdq
+       idiv     ebx
+ ;     mov      .dx12,eax
+       push      eax
+       mov     ax,.z2
+       sub     ax,.z1
+       cwde
+       shl     eax,CATMULL_SHIFT
+       cdq
+       idiv    ebx
+       push    eax
+       mov      ax,word[.b_x2]
+       sub      ax,word[.b_x1]
+       cwde
+       shl      eax,ROUND
+       cdq
+       idiv     ebx
+ ;     mov      .dbx12,eax
+       push      eax
+       mov      ax,word[.b_y2]
+       sub      ax,word[.b_y1]
+       cwde
+       shl      eax,ROUND
+       cdq
+       idiv     ebx
+ ;     mov      .dby12,eax
+       push      eax
+       mov      ax,word[.e_x2]
+       sub      ax,word[.e_x1]
+       cwde
+       shl      eax,ROUND
+       cdq
+       idiv     ebx
+ ;     mov      .dex12,eax
+       push      eax
+       mov      ax,word[.e_y2]
+       sub      ax,word[.e_y1]
+       cwde
+       shl      eax,ROUND
+       cdq
+       idiv     ebx
+ ;     mov      .dey12,eax
+       push      eax
+       mov      ax,word[.t_x2]
+       sub      ax,word[.t_x1]
+       cwde
+       shl      eax,ROUND
+       cdq
+       idiv     ebx
+ ;     mov      .dtx12,eax
+       push      eax
+       mov      ax,word[.t_y2]
+       sub      ax,word[.t_y1]
+       cwde
+       shl      eax,ROUND
+       cdq
+       idiv     ebx
+ ;     mov      .dty12,eax
+       push      eax
+end if
+   .bt_dx12_done:
+       mov      bx,.y3       ; calc delta13
+       sub      bx,.y1
+       jnz      .bt_dx13_make
+       mov      ecx,8
+       xor      edx,edx
+     @@:
+       push     edx   ;dword 0
+       loop     @b
+       jmp      .bt_dx13_done
+ .bt_dx13_make:
+       movsx    ebx,bx
+if Ext>=SSE
+       sub       esp,32
+   ;    mov       eax,256
+       cvtsi2ss  xmm4,[i255d]
+       cvtsi2ss  xmm3,ebx            ;rcps
+       divss     xmm3,xmm4
+       shufps    xmm3,xmm3,0
+       movd      mm0,[.b_x1]
+       movd      mm1,[.b_x3]
+       movd      mm2,[.e_x1]
+       movd      mm3,[.e_x3]
+       pxor       mm4,mm4
+       punpcklwd  mm0,mm4
+       punpcklwd  mm1,mm4
+       punpcklwd  mm2,mm4
+       punpcklwd  mm3,mm4
+       psubd      mm1,mm0
+       psubd      mm3,mm2
+       cvtpi2ps  xmm1,mm1
+       movlhps   xmm1,xmm1
+       cvtpi2ps  xmm1,mm3
+       divps     xmm1,xmm3   ;xmm1--> | dby | dbx | dey | dex |
+       shufps    xmm1,xmm1,10110001b
+                             ;xmm1--> | dbx | dby | dex | dey |
+;1       movups    .dey13,xmm1
+       cvtps2pi  mm0,xmm1 ;mm0,xmm1          ; mm0 -> 2 delta dwords
+       movhlps   xmm1,xmm1
+       cvtps2pi  mm1,xmm1 ;mm1,xmm1
+       movq      .dey13,mm0
+       movq      .dby13,mm1
+       mov       ax,.z3
+       sub       ax,.z1
+       cwde
+       mov      dx,.x3
+       sub      dx,.x1
+       movsx    edx,dx
+       movd       mm2,[.t_x1]
+       movd       mm3,[.t_x3]
+       punpcklwd  mm2,mm4
+       punpcklwd  mm3,mm4
+       psubd      mm3,mm2
+       cvtsi2ss  xmm1,eax
+       movlhps   xmm1,xmm1
+       cvtsi2ss  xmm1,edx
+       shufps    xmm1,xmm1,00101111b
+       cvtpi2ps  xmm1,mm3
+       divps     xmm1,xmm3   ; xmm1--> | dx | dz | dty | dtx |
+       shufps    xmm1,xmm1,11100001b
+                             ; xmm1--> | dx | dz | dtx | dty |
+;1       movlps    .dty13,xmm1
+;1       movhps    .dz13,xmm1
+       cvtps2pi  mm0,xmm1    ; mm0 -> 2 delta dwords  | dtx | dty |
+       movhlps   xmm1,xmm1
+       cvtps2pi  mm1,xmm1
+       movq      .dty13,mm0
+       movq      .dz13,mm1
+else
+       mov      ax,.x3
+       sub      ax,.x1
+       cwde
+       shl      eax,ROUND
+       cdq
+       idiv     ebx
+ ;     mov      .dx13,eax
+       push      eax
+       mov     ax,.z3
+       sub     ax,.z1
+       cwde
+       shl     eax,CATMULL_SHIFT
+       cdq
+       idiv    ebx
+  ;    mov    .dz13,eax
+       push    eax
+       mov      ax,word[.b_x3]
+       sub      ax,word[.b_x1]
+       cwde
+       shl      eax,ROUND
+       cdq
+       idiv     ebx
+ ;     mov      .dbx13,eax
+       push      eax
+       mov      ax,word[.b_y3]
+       sub      ax,word[.b_y1]
+       cwde
+       shl      eax,ROUND
+       cdq
+       idiv     ebx
+ ;     mov      .dby13,eax
+       push      eax
+       mov      ax,word[.e_x3]
+       sub      ax,word[.e_x1]
+       cwde
+       shl      eax,ROUND
+       cdq
+       idiv     ebx
+ ;     mov      .dex13,eax
+       push      eax
+       mov      ax,word[.e_y3]
+       sub      ax,word[.e_y1]
+       cwde
+       shl      eax,ROUND
+       cdq
+       idiv     ebx
+ ;     mov      .dey13,eax
+       push      eax
+       mov      ax,word[.t_x3]
+       sub      ax,word[.t_x1]
+       cwde
+       shl      eax,ROUND
+       cdq
+       idiv     ebx
+ ;     mov      .dtx13,eax
+       push      eax
+       mov      ax,word[.t_y3]
+       sub      ax,word[.t_y1]
+       cwde
+       shl      eax,ROUND
+       cdq
+       idiv     ebx
+ ;     mov      .dty13,eax
+       push      eax
+end if
+   .bt_dx13_done:
+       mov      bx,.y3       ; calc delta23
+       sub      bx,.y2
+       jnz      .bt_dx23_make
+       mov      ecx,8
+       xor      edx,edx
+     @@:
+       push     edx   ;dword 0
+       loop     @b
+       jmp      .bt_dx23_done
+ .bt_dx23_make:
+       movsx    ebx,bx
+if Ext>=SSE
+       sub       esp,32
+   ;    mov       eax,256
+       cvtsi2ss  xmm4,[i255d]
+       cvtsi2ss  xmm3,ebx            ;rcps
+       divss     xmm3,xmm4
+       shufps    xmm3,xmm3,0
+       movd      mm0,[.b_x2]
+       movd      mm1,[.b_x3]
+       movd      mm2,[.e_x2]
+       movd      mm3,[.e_x3]
+       pxor       mm4,mm4
+       punpcklwd  mm0,mm4
+       punpcklwd  mm1,mm4
+       punpcklwd  mm2,mm4
+       punpcklwd  mm3,mm4
+       psubd      mm1,mm0
+       psubd      mm3,mm2
+       cvtpi2ps  xmm1,mm1
+       movlhps   xmm1,xmm1
+       cvtpi2ps  xmm1,mm3
+       divps     xmm1,xmm3   ;xmm1--> | dby | dbx | dey | dex |
+       shufps    xmm1,xmm1,10110001b
+                             ;xmm1--> | dbx | dby | dex | dey |
+;1       movups    .dey23,xmm1
+       cvtps2pi  mm0,xmm1 ;mm0,xmm1          ; mm0 -> 2 delta dwords
+       movhlps   xmm1,xmm1
+       cvtps2pi  mm1,xmm1 ;mm1,xmm1
+       movq      .dey23,mm0
+       movq      .dby23,mm1
+       mov       ax,.z3
+       sub       ax,.z2
+       cwde
+       mov      dx,.x3
+       sub      dx,.x2
+       movsx    edx,dx
+       movd       mm2,[.t_x2]
+       movd       mm3,[.t_x3]
+       punpcklwd  mm2,mm4
+       punpcklwd  mm3,mm4
+       psubd      mm3,mm2
+       cvtsi2ss  xmm1,eax
+       movlhps   xmm1,xmm1
+       cvtsi2ss  xmm1,edx
+       shufps    xmm1,xmm1,00101111b
+       cvtpi2ps  xmm1,mm3
+       divps     xmm1,xmm3   ; xmm1--> | dx | dz | dty | dtx |
+       shufps    xmm1,xmm1,11100001b
+                            ; xmm1--> | dx | dz | dtx | dty |
+;       movlps    .dty23,xmm1
+;       movhps    .dz23,xmm1
+       cvtps2pi  mm0,xmm1    ; mm0 -> 2 delta dwords  | dtx | dty |
+       movhlps   xmm1,xmm1
+       cvtps2pi  mm1,xmm1    ; mm1 --> 2 delta dwords | dx  |  dz |
+       movq      .dty23,mm0
+       movq      .dz23,mm1
+else
+       mov      ax,.x3
+       sub      ax,.x2
+       cwde
+       shl      eax,ROUND
+       cdq
+       idiv     ebx
+ ;     mov      .dx23,eax
+       push      eax
+       mov     ax,.z3
+       sub     ax,.z2
+       cwde
+       shl     eax,CATMULL_SHIFT
+       cdq
+       idiv    ebx
+     ; mov     .dz23,eax
+       push    eax
+       mov      ax,word[.b_x3]
+       sub      ax,word[.b_x2]
+       cwde
+       shl      eax,ROUND
+       cdq
+       idiv     ebx
+ ;     mov      .dbx23,eax
+       push      eax
+       mov      ax,word[.b_y3]
+       sub      ax,word[.b_y2]
+       cwde
+       shl      eax,ROUND
+       cdq
+       idiv     ebx
+ ;     mov      .dby23,eax
+       push      eax
+       mov      ax,word[.e_x3]
+       sub      ax,word[.e_x2]
+       cwde
+       shl      eax,ROUND
+       cdq
+       idiv     ebx
+ ;     mov      .dex23,eax
+       push      eax
+       mov      ax,word[.e_y3]
+       sub      ax,word[.e_y2]
+       cwde
+       shl      eax,ROUND
+       cdq
+       idiv     ebx
+ ;     mov      .dey23,eax
+       push      eax
+       mov      ax,word[.t_x3]
+       sub      ax,word[.t_x2]
+       cwde
+       shl      eax,ROUND
+       cdq
+       idiv     ebx
+ ;     mov      .dtx23,eax
+       push      eax
+       mov      ax,word[.t_y3]
+       sub      ax,word[.t_y2]
+       cwde
+       shl      eax,ROUND
+       cdq
+       idiv     ebx
+ ;     mov      .dty23,eax
+       push      eax
+end if
+      ;  sub     esp,40
+   .bt_dx23_done:
+       sub       esp,64
+       movsx    eax,.x1
+       shl      eax,ROUND
+       mov      .cx1,eax
+       mov      .cx2,eax
+  ;     push     eax
+  ;     push     eax
+       movsx    ebx,word[.b_x1]
+       shl      ebx,ROUND
+       mov      .cbx1,ebx
+       mov      .cbx2,ebx
+      ; push     ebx
+      ; push     ebx
+       movsx    ecx,word[.b_y1]
+       shl      ecx,ROUND
+       mov      .cby1,ecx
+       mov      .cby2,ecx
+      ; push     ecx
+      ; push     ecx
+       movsx    edx,word[.e_x1]
+       shl      edx,ROUND
+       mov      .cex1,edx
+       mov      .cex2,edx
+    ;   push     edx
+    ;   push     edx
+       movsx    eax,word[.e_y1]
+       shl      eax,ROUND
+       mov      .cey1,eax
+       mov      .cey2,eax
+    ;   push     eax
+    ;   push     eax
+       movsx    ebx,.z1
+       shl      ebx,CATMULL_SHIFT
+       mov      .cz1,ebx
+       mov      .cz2,ebx
+   ;    push     ebx
+   ;    push     ebx
+      ; sub      esp,16
+       movsx    ecx,word[.t_x1]
+       shl      ecx,ROUND
+       mov      .ctx1,ecx
+       mov      .ctx2,ecx
+       ;push     ecx
+       ;push     ecx
+       movsx    edx,word[.t_y1]
+       shl      edx,ROUND
+       mov      .cty1,edx
+       mov      .cty2,edx
+      ; push     edx
+      ; push     edx
+if Ext >= SSE2
+       movups  xmm0,.cby1
+       movups  xmm1,.cty1
+       movups  xmm2,.cby2
+       movups  xmm3,.cty2
+       movups  xmm4,.dby13
+       movups  xmm5,.dty13
+       movups  xmm6,.dby12
+       movups  xmm7,.dty12
+       .scby1  equ [edi]
+       .scty1  equ [edi+16]
+       .scby2  equ [edi+32]
+       .scty2  equ [edi+48]
+       .sdby13 equ [edi+64]
+       .sdty13 equ [edi+80]
+       .sdby12 equ [edi+96]
+       .sdty12 equ [edi+128]
+       push    edi
+       mov     edi,sse_repository
+       movaps  .scby1,xmm0
+       movaps  .scty1,xmm1
+       movaps  .scby2,xmm2
+       movaps  .scty2,xmm3
+       movaps  .sdby13,xmm4
+       movaps  .sdty13,xmm5
+       movaps  .sdby12,xmm6
+       movaps  .sdty12,xmm7
+       pop     edi
+end if
+       movsx    ecx,.y1
+       cmp      cx,.y2
+       jge      .loop12_done
+  .loop12:
+;if Ext >= SSE2
+;       fxsave  [sse_repository]
+;end if
+       call     .call_line
+if Ext >= SSE2
+;       fxrstor [sse_repository]
+       movups  xmm0,.cby1
+       movups  xmm1,.cty1
+       movups  xmm2,.cby2
+       movups  xmm3,.cty2
+    ;   movups  xmm4,.dby13
+    ;   movups  xmm5,.dty13
+    ;   movups  xmm6,.dby12
+    ;   movups  xmm7,.dty12
+    ;   paddd   xmm0,xmm4
+    ;   paddd   xmm1,xmm5
+    ;   paddd   xmm2,xmm6
+    ;   paddd   xmm3,xmm7
+       push    edi
+       mov     edi,sse_repository
+       paddd   xmm0,.sdby13
+       paddd   xmm1,.sdty13
+       paddd   xmm2,.sdby12
+       paddd   xmm3,.sdty12
+       pop     edi
+       movups  .cby1,xmm0
+       movups  .cty1,xmm1
+       movups  .cby2,xmm2
+       movups  .cty2,xmm3
+end if
+if (Ext = MMX) | (Ext = SSE)
+       movq     mm0,.cby2
+       movq     mm1,.cby1
+       movq     mm2,.cey2
+       movq     mm3,.cey1
+       movq     mm4,.cty1
+       movq     mm5,.cty2
+       movq     mm6,.cz1
+       movq     mm7,.cz2
+       paddd    mm0,.dby12
+       paddd    mm1,.dby13
+       paddd    mm2,.dey12
+       paddd    mm3,.dey13
+       paddd    mm4,.dty13
+       paddd    mm5,.dty12
+       paddd    mm6,.dz13
+       paddd    mm7,.dz12
+       movq     .cby2,mm0
+       movq     .cby1,mm1
+       movq     .cey1,mm3
+       movq     .cey2,mm2
+       movq     .cty1,mm4
+       movq     .cty2,mm5
+       movq     .cz1,mm6
+       movq     .cz2,mm7
+end if
+if Ext = NON
+       mov      edx,.dbx13
+       add      .cbx1,edx
+       mov      eax,.dbx12
+       add      .cbx2,eax
+       mov      ebx,.dby13
+       add      .cby1,ebx
+       mov      edx,.dby12
+       add      .cby2,edx
+       mov      eax,.dex13
+       add      .cex1,eax
+       mov      ebx,.dex12
+       add      .cex2,ebx
+       mov      edx,.dey13
+       add      .cey1,edx
+       mov      eax,.dey12
+       add      .cey2,eax
+       mov      eax,.dtx13
+       add      .ctx1,eax
+       mov      ebx,.dtx12
+       add      .ctx2,ebx
+       mov      edx,.dty13
+       add      .cty1,edx
+       mov      eax,.dty12
+       add      .cty2,eax
+       mov      eax,.dx13
+       add      .cx1,eax
+       mov      ebx,.dx12
+       add      .cx2,ebx
+       mov      ebx,.dz13
+       add      .cz1,ebx
+       mov      edx,.dz12
+       add      .cz2,edx
+end if
+       inc      ecx
+       cmp      cx,.y2
+       jl       .loop12
+    .loop12_done:
+       movsx    ecx,.y2
+       cmp      cx,.y3
+       jge      .loop23_done
+       movsx    eax,.z2
+       shl      eax,CATMULL_SHIFT
+       mov      .cz2,eax
+       movsx    ebx,.x2
+       shl      ebx,ROUND
+       mov      .cx2,ebx
+       movzx    edx,word[.b_x2]
+       shl      edx,ROUND
+       mov      .cbx2,edx
+       movzx    eax,word[.b_y2]
+       shl      eax,ROUND
+       mov      .cby2,eax
+       movzx    ebx,word[.e_x2]
+       shl      ebx,ROUND
+       mov      .cex2,ebx
+       movzx    edx,word[.e_y2]
+       shl      edx,ROUND
+       mov      .cey2,edx
+       movzx    eax,word[.t_x2]
+       shl      eax,ROUND
+       mov      .ctx2,eax
+       movzx    ebx,word[.t_y2]
+       shl      ebx,ROUND
+       mov      .cty2,ebx
+if Ext >= SSE2
+       movups  xmm2,.cby2
+       movups  xmm3,.cty2
+   ;    movups  xmm4,.dby13
+   ;    movups  xmm5,.dty13
+       movups  xmm6,.dby23
+       movups  xmm7,.dty23
+;       .scby1  equ [edi]
+;       .scty1  equ [edi+16]
+;       .scby2  equ [edi+32]
+;       .scty2  equ [edi+48]
+;       .sdby13 equ [edi+64]
+;       .sdty13 equ [edi+80]
+       .sdby23 equ [edi+160]
+       .sdty23 equ [edi+192]
+       push    edi
+       mov     edi,sse_repository
+;       movaps  .scby1,xmm0
+;       movaps  .scty1,xmm1
+       movaps  .scby2,xmm2
+       movaps  .scty2,xmm3
+;       movaps  .sdby13,xmm4
+;       movaps  .sdty13,xmm5
+       movaps  .sdby23,xmm6
+       movaps  .sdty23,xmm7
+       pop     edi
+end if
+     .loop23:
+;if Ext >= SSE2
+;       fxsave  [sse_repository]
+;end if
+       call     .call_line
+if Ext >= SSE2
+       movups  xmm0,.cby1
+       movups  xmm1,.cty1
+       movups  xmm2,.cby2
+       movups  xmm3,.cty2
+       push    edi
+       mov     edi,sse_repository
+       paddd   xmm0,.sdby13
+       paddd   xmm1,.sdty13
+       paddd   xmm2,.sdby23
+       paddd   xmm3,.sdty23
+       pop     edi
+       movups  .cby1,xmm0
+       movups  .cty1,xmm1
+       movups  .cby2,xmm2
+       movups  .cty2,xmm3
+;       fxrstor [sse_repository]
+;       movups  xmm0,.cby1
+;       movups  xmm1,.cty1
+;       movups  xmm2,.cby2
+;       movups  xmm3,.cty2
+;       movups  xmm4,.dby13
+;       movups  xmm5,.dty13
+;       movups  xmm6,.dby23
+;       movups  xmm7,.dty23
+;       paddd   xmm0,xmm4
+;       paddd   xmm1,xmm5
+;       paddd   xmm2,xmm6
+ ;      paddd   xmm3,xmm7
+ ;      movups  .cby1,xmm0
+ ;      movups  .cty1,xmm1
+ ;      movups  .cby2,xmm2
+ ;      movups  .cty2,xmm3
+;
+end if
+if (Ext = MMX) | (Ext = SSE)
+       movq     mm0,.cby2
+       movq     mm1,.cby1
+       movq     mm2,.cey2
+       movq     mm3,.cey1
+       movq     mm4,.cty1
+       movq     mm5,.cty2
+       movq     mm6,.cz1
+       movq     mm7,.cz2
+       paddd    mm0,.dby23
+       paddd    mm1,.dby13
+       paddd    mm2,.dey23
+       paddd    mm3,.dey13
+       paddd    mm4,.dty13
+       paddd    mm5,.dty23
+       paddd    mm6,.dz13
+       paddd    mm7,.dz23
+       movq     .cby2,mm0
+       movq     .cby1,mm1
+       movq     .cey2,mm2
+       movq     .cey1,mm3
+       movq     .cty1,mm4
+       movq     .cty2,mm5
+       movq     .cz1,mm6
+       movq     .cz2,mm7
+end if
+If Ext = NON
+       mov      edx,.dbx13
+       add      .cbx1,edx
+       mov      eax,.dbx23
+       add      .cbx2,eax
+       mov      ebx,.dby13
+       add      .cby1,ebx
+       mov      edx,.dby23
+       add      .cby2,edx
+       mov      eax,.dex13
+       add      .cex1,eax
+       mov      ebx,.dex23
+       add      .cex2,ebx
+       mov      edx,.dey13
+       add      .cey1,edx
+       mov      eax,.dey23
+       add      .cey2,eax
+       mov      eax,.dx13
+       add      .cx1,eax
+       mov      ebx,.dx23
+       add      .cx2,ebx
+       mov      ebx,.dz13
+       add      .cz1,ebx
+       mov      edx,.dz23
+       add      .cz2,edx
+       mov      eax,.dtx13
+       add      .ctx1,eax
+       mov      ebx,.dtx23
+       add      .ctx2,ebx
+       mov      edx,.dty13
+       add      .cty1,edx
+       mov      eax,.dty23
+       add      .cty2,eax
+end if
+       inc      ecx
+       cmp      cx,.y3
+       jl       .loop23
+    .loop23_done:
+       mov      esp,ebp
+ret   50
+.call_line:
+       pushad
+       ; xmm0= cby1,cbx1,cz1,cx1
+       ; xmm1= cty1,ctx1,cey1,cex1
+if Ext >= SSE2
+       sub      esp,8
+       shufps   xmm1,xmm1,10110001b
+       shufps   xmm3,xmm3,10110001b
+       movlps   [esp],xmm1
+else
+       push     dword .cty1
+       push     .ctx1
+end if
+       push     dword .cz1
+if Ext>=SSE2
+       sub      esp,8
+       movlps   [esp],xmm3
+else
+       push     dword .cty2
+       push     .ctx2
+end if
+       push     dword .cz2
+if Ext>=SSE2
+       sub      esp,32
+       movhps   [esp+24],xmm3
+       shufps   xmm2,xmm2,10110001b
+       movlps   [esp+16],xmm2
+       movhps   [esp+8],xmm1
+       shufps   xmm0,xmm0,10110001b
+       movlps   [esp],xmm0 ;================================
+else
+       push     dword .cey2
+       push     .cex2
+       push     dword .cby2
+       push     .cbx2
+       push     dword .cey1
+       push     .cex1
+       push     dword .cby1
+       push     .cbx1
+end if
+       push     .tex_ptr
+       push     .z_buff
+       push     .t_emap
+       push     .t_bmap
+       push     ecx
+       mov      eax,.cx1
+       sar      eax,ROUND
+       mov      ebx,.cx2
+       sar      ebx,ROUND
+       call     bump_tex_line_z
+       popad
+;end if
+ret
+bump_tex_line_z:
+;--------------in: eax - x1
+;--------------    ebx - x2
+;--------------    edi - pointer to screen buffer
+;stack - another parameters :
+.y      equ dword [ebp+4]
+.bmap   equ dword [ebp+8]        ; bump map pointer
+.emap   equ dword [ebp+12]       ; env map pointer
+.z_buff equ dword [ebp+16]       ; z buffer
+.tex_map equ dword [ebp+20]      ; texture pointer
+.bx1    equ  [ebp+24]   ;   ---
+.by1    equ  [ebp+28]  ;       |
+.ex1    equ  [ebp+32]  ;       |
+.ey1    equ  [ebp+36]  ;       |
+.bx2    equ  [ebp+40]  ;       |
+.by2    equ  [ebp+44]  ;       |>   b. map and e. map coords
+.ex2    equ  [ebp+48]  ;       |>   shifted shl ROUND
+.ey2    equ  [ebp+52]  ;   ---
+.z2     equ  [ebp+56]
+.tx2    equ  [ebp+60]
+.ty2    equ  [ebp+64]
+.z1     equ  [ebp+68]
+.tx1    equ  [ebp+72]
+.ty1    equ  [ebp+76]
+.x1     equ [ebp-4]
+.x2     equ [ebp-8]
+.dbx    equ [ebp-12]
+.dby    equ [ebp-16]
+.dex    equ [ebp-20]
+.dey    equ [ebp-24]
+.dz     equ [ebp-28]
+.dtx    equ [ebp-32]
+.dty    equ [ebp-36]
+.cbx    equ [ebp-40]
+.cby    equ [ebp-44]
+.cex    equ [ebp-48]
+.cey    equ [ebp-52]
+.cz     equ [ebp-56]
+.czbuff equ [ebp-60]
+.ctx    equ [ebp-64]
+.cty    equ [ebp-68]
+.c_scr  equ [ebp-72]
+.temp1  equ        ebp-80
+.temp2  equ        ebp-88
+.temp3  equ        ebp-76
+.temp4  equ        ebp-84
+.temp5  equ        ebp-92
+        mov     ebp,esp
+        mov     ecx,.y
+        or      ecx,ecx
+        jl      .bl_end
+        movzx   edx,word[size_y_var]
+        cmp     ecx,edx  ;SIZE_Y
+        jge     .bl_end
+        cmp     eax,ebx
+        jl      .bl_ok
+        je      .bl_end
+if Ext=NON
+        mov     edx,.bx1
+        xchg    edx,.bx2
+        mov     .bx1,edx
+        mov     edx,.by1
+        xchg    edx,.by2
+        mov     .by1,edx
+        mov     edx,.ex1
+        xchg    edx,.ex2
+        mov     .ex1,edx
+        mov     edx,.ey1
+        xchg    edx,.ey2
+        mov     .ey1,edx
+        mov     edx,.tx1
+        xchg    edx,.tx2
+        mov     .tx1,edx
+        mov     edx,.ty1
+        xchg    edx,.ty2
+        mov     .ty1,edx
+end if
+if Ext = MMX
+        movq    mm0,.bx1
+        movq    mm1,.bx2
+        movq    mm2,.ex1
+        movq    mm3,.ex2
+        movq    mm4,.tx1
+        movq    mm5,.tx2
+        movq    .bx2,mm0
+        movq    .bx1,mm1
+        movq    .ex1,mm3
+        movq    .ex2,mm2
+        movq    .tx1,mm5
+        movq    .tx2,mm4
+end if
+if Ext>=SSE
+        movups xmm0,.bx1
+        movups xmm1,.bx2
+        movups .bx1,xmm1
+        movups .bx2,xmm0
+        movq    mm0,.tx1
+        movq    mm1,.tx2
+        movq    .tx1,mm1
+        movq    .tx2,mm0
+end if
+;if Ext>=SSE2
+;        movaps  xmm4,xmm0
+;        movaps  xmm0,xmm2
+;        movaps  xmm2,xmm4
+;        movaps  xmm5,xmm1
+;        movaps  xmm1,xmm3
+;        movaps  xmm3,xmm5
+;else
+        xchg    eax,ebx
+        mov     edx,.z1
+        xchg    edx,.z2
+        mov     .z1,edx
+;end if
+  .bl_ok:
+;if Ext >= SSE2
+;        shufps  xmm0,xmm0,11100001b
+;        shufps  xmm2,xmm2,11100001b
+;        movlps  .bx1,xmm0
+;        movlps  .bx2,xmm2
+;        shufps  xmm0,xmm0,00011011b
+;        shufps  xmm2,xmm2,00011011b
+;        movd    eax,xmm0
+;        movd    ebx,xmm2
+;        shufps  xmm0,xmm0,11000110b
+;        shufps  xmm2,xmm2,11000110b
+;        movd    .z1,xmm0
+;        movd    .z2,xmm2
+;        shufps  xmm1,xmm1,10110001b
+;        shufps  xmm3,xmm3,10110001b
+;        movlps  .ex1,xmm1
+;        movlps  .ex2,xmm2
+;        movhps  .tx1,xmm1
+;        movhps  .tx2,xmm2
+;        xchg    eax,ebx
+;        mov     edx,.z1
+;        xchg    edx,.z2
+;        mov     .z1,edx
+;end if
+        push    eax
+        push    ebx           ;store x1, x2
+        movzx   ebx,word[size_x_var]
+    ;    mov     eax,.x1
+        cmp     dword .x1,ebx  ;dword .x1,SIZE_X
+        jge     .bl_end
+        cmp     dword .x2,0
+        jle     .bl_end
+        mov     ebx,.x2
+        sub     ebx,.x1
+if Ext>=SSE
+       sub       esp,28
+       cvtsi2ss  xmm3,ebx            ;rcps
+       shufps    xmm3,xmm3,0
+; float using SSE variant  ::-->
+;       movups    xmm0,.bx1  ; new
+;       movups    xmm1,.bx2  ; new
+       cvtpi2ps  xmm0,.bx1 ;mm0    ; variant fixed point
+       movlhps   xmm0,xmm0
+       cvtpi2ps  xmm0,.ex1 ;mm2
+       cvtpi2ps  xmm1,.bx2 ;mm1
+       movlhps   xmm1,xmm1
+       cvtpi2ps  xmm1,.ex2 ;mm3
+       subps     xmm1,xmm0
+       divps     xmm1,xmm3
+       shufps    xmm1,xmm1,10110001b
+;       movups    .dey,xmm1  ; new
+       cvtps2pi  mm0,xmm1          ; mm0 -> 2 delta dwords
+       movhlps   xmm1,xmm1
+       cvtps2pi  mm1,xmm1
+       movq      .dey,mm0
+       movq      .dby,mm1
+       movd      mm2,.z1
+       movd      mm3,.z2
+       cvtpi2ps  xmm0,.tx1 ;mm0
+       movlhps   xmm0,xmm0
+       cvtpi2ps  xmm0,mm2
+       cvtpi2ps  xmm1,.tx2 ;mm1
+       movlhps   xmm1,xmm1
+       cvtpi2ps  xmm1,mm3
+;       movups    xmm0,,z1  ; new
+;       movups    xmm1,.z2  ; new
+       subps     xmm1,xmm0
+       divps     xmm1,xmm3
+;       movups    .dz,xmm1  ;new
+       shufps    xmm1,xmm1,10110100b
+       cvtps2pi  mm0,xmm1          ; mm0 -> 2 delta dwords
+       movhlps   xmm1,xmm1
+       cvtps2pi  mm1,xmm1
+       movd      .dz,mm0
+       movq      .dty,mm1
+else
+        mov     eax,.bx2       ; calc .dbx
+        sub     eax,.bx1
+        cdq
+        idiv    ebx
+        push    eax
+        mov     eax,.by2       ; calc .dby
+        sub     eax,.by1
+        cdq
+        idiv    ebx
+        push    eax
+        mov     eax,.ex2       ; calc .dex
+        sub     eax,.ex1
+        cdq
+        idiv    ebx
+        push    eax
+        mov     eax,.ey2       ; calc .dey
+        sub     eax,.ey1
+        cdq
+        idiv    ebx
+        push    eax
+        mov     eax,.z2        ; calc .dz
+        sub     eax,.z1
+        cdq
+        idiv    ebx
+        push    eax
+        mov     eax,.tx2       ; calc .dtx
+        sub     eax,.tx1
+        cdq
+        idiv    ebx
+        push    eax
+        mov     eax,.ty2       ; calc .dty
+        sub     eax,.ty1
+        cdq
+        idiv    ebx
+        push    eax
+end if
+        cmp     dword .x1,0         ; set correctly begin variable
+        jge     @f            ; CLIPPING ON FUNCTION
+                              ; cutting triangle exceedes screen
+        mov     ebx,.x1
+        neg     ebx
+;if Ext >= SSE
+;        cvtsi2ss xmm0,ebx
+;        shufps   xmm0,xmm0,0
+;        movups   xmm1,.dey
+;        mulps    xmm1,xmm0
+;        shufps   xmm1,xmm1,00011011b
+;        movups   xmm2,.bx1
+;        addps    xmm2,xmm1
+;        movups   .bx1,xmm2
+        mov     eax,.dz
+        imul    ebx           ; eax = .dz * abs(.x1)
+        add     .z1,eax
+        mov     dword .x1,0
+        mov     eax,.dbx
+        imul    ebx
+        add    .bx1,eax
+        mov     eax,.dby
+        imul    ebx
+        add     .by1,eax
+        mov     eax,.dex
+        imul    ebx
+        add     .ex1,eax
+        mov     eax,.dey
+        imul    ebx
+        add     .ey1,eax
+        mov     eax,.dtx
+        imul    ebx
+        add     .tx1,eax
+        mov     eax,.dty
+        imul    ebx
+        add     .ty1,eax
+      @@:
+   ;     mov     ebx,.x2
+        movzx   eax,word[size_x_var]
+       ; cmp     dword .x2,SIZE_X
+        cmp     dword .x2,eax  ; eax,ebx
+        jl      @f
+        mov     dword .x2,eax  ;SIZE_X
+      @@:
+        movzx   eax,word[size_x_var]  ;SIZE_X       ;calc memory begin in buffers
+        mul     .y
+        add     eax,.x1
+        lea     esi,[4*eax]
+        add     esi,.z_buff       ; z-buffer filled with dd variables
+        lea     eax,[eax*3]
+        add     edi,eax
+        mov     ecx,.x2
+        sub     ecx,.x1
+        ; init current variables
+        push    dword .bx1   ; current b, e and t shifted shl ROUND   .cbx
+        push    dword .by1                                         ;  .cby
+        push    dword .ex1                                         ;  .cex
+        push    dword .ey1                                         ;  .cey
+        push    dword .z1    ; current z shl CATMULL_SHIFT         ; .cz
+        push    esi                                          ; .czbuff
+        push    dword .tx1      ;         .ctx
+        push    dword .ty1      ;         .cty
+        push    edi       ;         .c_scr
+if Ext = SSE2
+        mov    eax,TEXTURE_SIZE
+        movd   xmm1,eax
+        shufps xmm1,xmm1,0
+        push   dword  TEX_X
+        push   dword  -TEX_X
+        push   dword  1
+        push   dword  -1
+        movups xmm2,[esp]
+        movd   xmm3,.bmap
+        shufps xmm3,xmm3,0
+end if
+if Ext>=MMX
+        movq    mm7,.cty
+        movq    mm6,.cby
+        movq    mm5,.cey
+;        movq    mm4,.dtyq
+;        movq    mm3,.dbyq
+end if
+     .draw:
+    ; if TEX = SHIFTING   ;bump drawing only in shifting mode
+        mov     esi,.czbuff      ; .czbuff current address in buffer
+        mov     ebx,.cz          ; .cz - cur z position
+        cmp     ebx,dword[esi]
+        jge     .skip
+if Ext=NON
+        mov     eax,.cby
+        shr     eax,ROUND
+        mov     esi,.cbx
+        shr     esi,ROUND
+else
+        movq    mm1,mm6
+        psrld   mm1,ROUND
+        movd    eax,mm1
+        psrlq   mm1,32
+        movd    esi,mm1
+end if
+        shl     eax,TEX_SHIFT
+        add     esi,eax         ;-  ; esi - current bump map index
+if Ext = SSE2
+        movd    xmm0,esi
+        shufps  xmm0,xmm0,0
+        paddd   xmm0,xmm2
+        pand    xmm0,xmm1
+        paddd   xmm0,xmm3
+        movd    ebx,xmm0
+        movzx   eax,byte[ebx]
+;
+;        shufps  xmm0,xmm0,11100001b
+        psrldq  xmm0,4
+        movd    ebx,xmm0
+        movzx   ebx,byte[ebx]
+        sub     eax,ebx
+;
+;        shufps  xmm0,xmm0,11111110b
+        psrldq  xmm0,4
+        movd    ebx,xmm0
+        movzx   edx, byte [ebx]
+;
+;        shufps  xmm0,xmm0,11111111b
+        psrldq  xmm0,4
+        movd    ebx,xmm0
+        movzx   ebx, byte [ebx]
+        sub     edx,ebx
+;
+else
+;        mov     ebx,esi
+;        dec     ebx
+        lea     ebx,[esi-1]
+        and     ebx,TEXTURE_SIZE
+        add     ebx,.bmap
+        movzx   eax,byte [ebx]
+;        mov     ebx,esi
+;        inc     ebx
+        lea     ebx,[esi+1]
+        and     ebx,TEXTURE_SIZE
+        add     ebx,.bmap
+        movzx   ebx,byte [ebx]
+        sub     eax,ebx
+;        mov     ebx,esi
+;        sub     ebx,TEX_X
+        lea     ebx,[esi-TEX_X]
+        and     ebx,TEXTURE_SIZE
+        add     ebx,.bmap
+        movzx   edx,byte [ebx]
+;        mov     ebx,esi
+;        add     ebx,TEX_X
+        lea     ebx,[esi+TEX_X]
+        and     ebx,TEXTURE_SIZE
+        add     ebx,.bmap
+        movzx   ebx,byte [ebx]
+        sub     edx,ebx
+end if
+     ;  eax - horizontal sub    modificated x coord
+     ;  edx - vertical   sub    modificated y coord
+if Ext=NON
+        mov     ebx,.cex       ;.cex - current env map X
+        shr     ebx,ROUND
+        add     eax,ebx
+        mov     ebx,.cey       ;.cey - current  env map y
+        shr     ebx,ROUND
+        add     edx,ebx
+else
+        movq    mm1,mm5        ; mm5 - copy of cur env coords
+        psrld   mm1,ROUND
+        movd    ebx,mm1
+        psrlq   mm1,32
+        add     eax,ebx
+        movd    ebx,mm1
+        add     edx,ebx
+;        movq    qword[.temp1],mm3
+;        add     eax,dword [.temp1]
+;        add     edx,dword [.temp1+4]
+end if
+        or      eax,eax
+        jl      .black
+        cmp     eax,TEX_X
+        jg      .black
+        or      edx,edx
+        jl      .black
+        cmp     edx,TEX_Y
+        jg      .black
+        shl     edx,TEX_SHIFT   ; zaburzenie w emapie = zaburzenie w teksturze
+        add     edx,eax         ; proponuje nie stawiac czarnego pixela tylko
+        lea     esi,[edx*3]     ; niezaburzony.
+        add     esi,.emap       ;
+        lodsd
+if Ext=NON
+        mov     edx,.cty
+        shr     edx,ROUND  ; sar
+        mov     edi,.ctx
+        shr     edi,ROUND  ; sar
+else
+        movq    mm1,mm7
+        psrld   mm1,ROUND
+        movd    edx,mm1
+        psrlq   mm1,32
+        movd    edi,mm1
+end if
+        shl     edx,TEX_SHIFT
+        add     edi,edx
+        and     edi,TEXTURE_SIZE
+        lea     esi,[edi*3]
+        add     esi,.tex_map
+if Ext=NON
+        mov     edx,eax
+        lodsd
+        push    ax
+        mul     dl
+        mov     dl,ah
+        pop     ax
+        shr     ax,8
+        mul     dh
+        mov     al,dl
+        mov     edi,.c_scr
+        stosw
+        shr     edx,16
+        shr     eax,16
+        mul     dl
+        shr     ax,8
+        stosb
+else
+        movd       mm0,eax
+        pxor       mm1,mm1
+        punpcklbw  mm0,mm1
+        movd       mm2,[esi]
+        punpcklbw  mm2,mm1
+        pmullw     mm0,mm2
+        psrlw      mm0,8
+        packuswb   mm0,mm1
+        mov        edi,.c_scr
+        movd       [edi],mm0
+end if
+        jmp     .actual_zbuff   ; actualize z buffer
+     @@:
+     .black:
+        xor     eax,eax
+        mov     edi,.c_scr
+        stosd
+     .actual_zbuff:
+        mov     eax,.cz
+        mov     edi,.czbuff
+        stosd
+      .skip:
+        add     dword .czbuff,4
+        add     dword .c_scr,3
+if Ext=NON
+        mov     eax,.dbx
+        add     .cbx,eax
+        mov     ebx,.dby
+        add     .cby,ebx
+        mov     edx,.dex
+        add     .cex,edx
+        mov     eax,.dey
+        add     .cey,eax
+        mov     ebx,.dtx
+        add     .ctx,ebx
+        mov     edx,.dty
+        add     .cty,edx
+else
+        paddd   mm7,.dty
+        paddd   mm6,.dby
+        paddd   mm5,.dey
+end if
+        mov     eax,.dz
+        add     .cz,eax
+        dec     ecx
+        jnz     .draw
+  .bl_end:
+        mov     esp,ebp
+ret 76
+;Ext = MMX
+;     else
+;        movq    mm5, qword[.temp1]  ;-
+;        paddd   mm5, qword[.temp5]  ; .temp5 == low dword = TEX_X, high dword = -TEX_X
+;        pand    mm5, qword[.temp3]  ; .temp3 == low = high dword = TEX_SIZE
+;        paddd   mm5, qword[.temp4]  ; .temp4 == low = high dword = offset .bmap
+;        movd    ebx,mm5
+;        psrlq   mm5,32
+;     end if

 /programs/demos/view3ds/data.inc
 ,5 → 1,7
                    ; DATA AREA  ************************************
+      if Ext > SSE2
+        isSSE3          db      1
+      end if
         i3              dw      3
         i6              dd      6
         i12             dd      12
 ,6 → 8,7
         i36             dd      36
         i256            dw      256
         i255d           dd      255
+        f1:
         dot_max         dd      1.0      ; dot product max and min
         dot_min         dd      0.0
         env_const       dd      1.05
 ,7 → 28,7
         y_offset      dw SIZE_Y / 2
         z_offset      dw 0
         rsscale       dd 175.0           ; next real scale
-        vect_x        dw SIZE_X / 2
+        vect_x:        dw SIZE_X / 2
         vect_y        dw SIZE_Y / 2
         vect_z        dw 0
                 size_y_var:
 ,9 → 113,9
                dd ?
                db 7
-               db 'catmull   '
+               db 'ray shadow'
                db 2
-catmull_flag   db 1
+ray_shd_flag   db 0
                dd onoff_f
                db 8
 ,7 → 168,7
                db 16
                db 'fire      '
-               db 3
+               db 2
 fire_flag      db 0
                dd blur_f
 ,7 → 353,7
       if Ext=SSE3
         db   ' (SSE3)'
       end if
-        db   ' 0.073',0
+        db   ' 0.074',0
     labellen:
         STRdata db '-1        '
      lab_vert:
 ,8 → 428,43
+;if Ext >= SSE3
+align 16
+   point_light_coords:
+              dd 50.0
+              dd 50.0
+              dd -215.0
+              dd 0.0
-align 16
+              dd 815.0
+              dd 815.0
+              dd -215.0
+              dd 0.0
+              dd 1500.0
+              dd 1500.0
+              dd -215.0
+              dd 0.0
+if 0
+    aabb1:
+        .0              dd 1.0,1.0,1.0,0
+        .1              dd -1.0,1.0,1.0,0
+        .2              dd 1.0,-1.0,1.0,0
+        .3              dd -1.0,-1.0,1.0,0
+        .4              dd 1.0,1.0,-1.0,0
+        .5              dd -1.0,1.0,-1.0,0
+        .6              dd 1.0,-1.0,-1.0,0
+        .7              dd -1.0,-1.0,-1.0,0
+end if
+    sign_mask:
+              times 4 dd 0x80000000
+    f05xz:    dd 0, 0,   - 1.0  ,0
+    sign_z:
+              dd -1,-1,0x7fffffff,0
     abs_mask:
            dd 0x7fffffff, 0x7fffffff, 0x7fffffff, 0x7fffffff
     emboss_bias:
 ,9 → 480,13
            times 4 dd 510.0
     the_one:
            times 4 dd 1.0
-     aprox  dd 0.0001
+     eps:  times 4 dd 0.00000
+     epsone      dd 1.0001
+     aprox       dd 0.0001
+     epsminus    dd -0.0001
         file_info:
                 dd      0
                 dd      0
 ,22 → 505,13
 workarea   rb 180
 EndFile   dd ?
 align 8
-        sinbeta dd      ?;+32
+        sinbeta dd      ?;
         cosbeta dd      ?
         xsub    dw      ?
-        zsub    dw      ?;+40
+        zsub    dw      ?
         ysub    dw      ?
-        xx1     dw      ?
-        yy1     dw      ?
-        zz1     dw      ?;+48    xx1 + 4
-        xx2     dw      ?
-        yy2     dw      ?
-        zz2     dw      ?      ; xx1 + 10
-        xx3     dw      ?;+56
-        yy3     dw      ?
-        zz3     dw      ?      ; xx1 + 16
         col1    dd      ?
         col2    dd      ?
         col3    dd      ?
 ,13 → 520,9
         points_count_var        dd ?  ;
         triangles_count_var     dd ?  ; dont change order
         edges_count             dd ?  ;
+        tex_points_ptr          dd ?
-        point_index1        dd ?   ;-\
-        point_index2        dd ?   ;  }  don't change order
-        point_index3        dd ?   ;-/
         temp_col            dw ?
-        temp1               dd ?   ; > dont change
-        temp2               dd ?   ; >  order
         high                dd ?
         rand_seed           dw ?
 align 8
 ,18 → 539,14
         matrix  rb 36
         cos_tab rd 360
         sin_tab rd 360
 align 16
+       lights_aligned:
+       lights_aligned_end = $ + 16 * 12
+                        rb 16 * 12
-    points_count = 180000/6*3
-    triangles_count = 180000 / 6  ;($-triangles)/6
 align 16
-        label trizdd dword
-        label trizdq qword
-        triangles_with_z rw triangles_count*4 + 2 ; triangles triple dw + z position
-align 16
         vectors rb 24
 align 16
         bumpmap         rb      TEXTURE_SIZE + 1
 align 16
 ,25 → 560,19
 align 16
         color_map       rb      (TEXTURE_SIZE +100) * 3
 align 16
-        tex_points      rb      points_count * 4  ; bump_map  and texture coords
-                                ; each point word x, word y
-align 16
-       lights_aligned:
-       lights_aligned_end = $ + 16 * 12
-                        rb 16 * 12
+ ;       tex_points      rb      points_count * 4  ; bump_map  and texture coords
+ ;                               ; each point word x, word y
+;align 16
+;       lights_aligned:
+;       lights_aligned_end = $ + 16 * 12
+;                        rb 16 * 12
 if Ext >= SSE2
         sse_repository  rb      1024
 end if
-  ;      SourceFile:             ; source file temporally in screen area
-  ;      workarea        dd      ?
-    ;    screen          rb      SIZE_X * SIZE_Y * 3   ; screen buffer
-;align 16
-     ;   Z_buffer        rb      SIZE_X * SIZE_Y * 4
         procinfo:
-                        rb     1024 ; process info
+                        rb     2048 ; process info
         I_Param         rb      256
         memStack:
         rb 2000

 /programs/demos/view3ds/flat_cat.inc
 ,399 → 1,399
-CATMULL_SHIFT equ 16
-flat_triangle_z:
-;  procedure drawing triangle with Z cordinate interpolation ------
-;  (Catmull  alghoritm)--------------------------------------------
-;  ----------------in - eax - x1 shl 16 + y1 ----------------------
-;  -------------------- ebx - x2 shl 16 + y2 ----------------------
-;  -------------------- ecx - x3 shl 16 + y3 ----------------------
-;  -------------------- edx - color 0x00RRGGBB --------------------
-;  -------------------- esi - pointer to Z-buffer -----------------
-;  -------------------- edi - pointer to screen buffer-------------
-;  -------------------- stack : z coordinates
-;  -------------------- Z-buffer : each z variable as dword
-;  -------------------- (Z coor. as word) shl CATMULL_SHIFT
-.z1     equ word[ebp+4]
-.z2     equ word[ebp+6]    ; each z coordinate as word integer
-.z3     equ word[ebp+8]
-.col    equ dword[ebp-4]
-.x1     equ word[ebp-6]
-.y1     equ word[ebp-8]
-.x2     equ word[ebp-10]
-.y2     equ word[ebp-12]
-.x3     equ word[ebp-14]
-.y3     equ word[ebp-16]
-.dx12   equ dword[ebp-20]
-;.dz12   equ dword[ebp-24]
-.dx13   equ dword[ebp-24]
-.dz13   equ dword[ebp-28]
-.dz12   equ dword[ebp-32]
-;.dz13   equ dword[ebp-32]
-.dx23   equ dword[ebp-36]
-.dz13M  equ      [ebp-40]
-.dz23   equ dword[ebp-44]
-.zz1    equ dword[ebp-48]
-.zz2    equ dword[ebp-52]
-.zz2M   equ qword[ebp-52]
-.dz12M  equ qword[ebp-32]
-.dz23M  equ qword[ebp-44]
-;if Ext>=MMX
-;        emms
-;end if
-        mov     ebp,esp
-        push    edx            ; store edx in variable .col
-      .sort2:
-        cmp     ax,bx
-        jle     .sort1
-        xchg    eax,ebx
-        mov     dx,.z1
-        xchg    dx,.z2
-        mov     .z1,dx
-      .sort1:
-        cmp     bx,cx
-        jle     .sort3
-        xchg    ebx,ecx
-        mov     dx,.z2
-        xchg    dx,.z3
-        mov     .z2,dx
-        jmp     .sort2
-      .sort3:
-        push    eax            ; store triangle coordinates in user friendly variables
-        push    ebx
-        push    ecx
-         mov      edx,80008000h  ; eax,ebx,ecx are ANDd together into edx which means that
-         and      edx,ebx        ; if *all* of them are negative a sign flag is raised
-         and      edx,ecx
-         and      edx,eax
-         test     edx,80008000h  ; Check both X&Y at once
-         jne      .ft_loop2_end
-       ;  cmp      ax,SIZE_Y
-       ;  jle      @f
-       ;  cmp      bx,SIZE_Y
-       ;  jle      @f
-       ;  cmp      cx,SIZE_Y
-       ;  jge      @f
-       ;  ror      eax,16
-       ;  ror      ebx,16
-       ;  ror      ecx,16
-       ;  cmp      ax,SIZE_X
-       ;  jle      @f
-       ;  cmp      bx,SIZE_X
-       ;  jle      @f
-       ;  cmp      cx,SIZE_X
-       ;  jle      @f
-       ;  jmp      .ft_loop2_end
-       ;@@:
-        sub     esp,52-12
-        mov     bx,.y2       ; calc delta 12
-        sub     bx,.y1
-        jnz     .ft_dx12_make
-        mov     .dx12,0
-        mov     .dz12,0
-        jmp     .ft_dx12_done
-   .ft_dx12_make:
-        mov     ax,.x2
-        sub     ax,.x1
-        cwde
-        movsx   ebx,bx
-        shl     eax,ROUND
-        cdq
-        idiv    ebx
-        mov     .dx12,eax
-        mov     ax,.z2
-        sub     ax,.z1
-        cwde
-        shl     eax,CATMULL_SHIFT
-        cdq
-        idiv    ebx
-        mov     .dz12,eax
-  .ft_dx12_done:
-        mov     bx,.y3       ; calc delta 13
-        sub     bx,.y1
-        jnz     .ft_dx13_make
-        mov     .dx13,0
-        mov     .dz13,0
-        mov     dword .dz13M,0
-        jmp     .ft_dx13_done
-   .ft_dx13_make:
-        mov     ax,.x3
-        sub     ax,.x1
-        cwde
-        movsx   ebx,bx
-        shl     eax,ROUND
-        cdq
-        idiv    ebx
-        mov     .dx13,eax
-        mov     ax,.z3
-        sub     ax,.z1
-        cwde
-        shl     eax,CATMULL_SHIFT
-        cdq
-        idiv    ebx
-        mov     .dz13,eax
-        mov     dword .dz13M,eax
-    .ft_dx13_done:
-        mov     bx,.y3       ; calc delta 23
-        sub     bx,.y2
-        jnz     .gt_dx23_make
-        mov     .dx23,0
-        mov     .dz23,0
-        jmp     .gt_dx23_done
-   .gt_dx23_make:
-        mov     ax,.x3
-        sub     ax,.x2
-        cwde
-        movsx   ebx,bx
-        shl     eax,ROUND
-        cdq
-        idiv    ebx
-        mov     .dx23,eax
-        mov     ax,.z3
-        sub     ax,.z2
-        cwde
-        shl     eax,CATMULL_SHIFT
-        cdq
-        idiv    ebx
-        mov     .dz23,eax
-    .gt_dx23_done:
-        movsx   edx,.z1
-        shl     edx,CATMULL_SHIFT
-        mov     .zz1,edx
-        mov     .zz2,edx
-        movsx   eax,.x1
-        shl     eax,ROUND    ; eax - x1
-        mov     ebx,eax      ; ebx - x2
-;if Ext>=MMX
-;        movq    mm0,.zz2M
-;end if
-        mov     cx,.y1
-        cmp     cx,.y2
-        jge     .ft_loop1_end
-     .ft_loop1:
-        pushad
-        push    .col
-        push    cx          ; y
-        sar     ebx,ROUND
-        push    bx          ; x2
-        sar     eax,ROUND
-        push    ax          ; x1
-;if Ext>=MMX
-;        sub     esp,8
-;        movq    [esp],mm0
-;else
-        push    .zz2        ; z2 shl CATMULL_SHIFT
-        push    .zz1        ; z1 shl CATMULL_SHIFT
-;end if
-        call    flat_line_z
-        popad
-        add     eax,.dx13
-        add     ebx,.dx12
-;if  Ext>=MMX
-;        paddd   mm0,.dz12M
-;else
-        mov     edx,.dz13
-        add     .zz1,edx
-        mov     edx,.dz12
-        add     .zz2,edx
-;end if
-        inc     cx
-        cmp     cx,.y2
-        jl      .ft_loop1
-    .ft_loop1_end:
-        movsx   edx,.z2
-        shl     edx,CATMULL_SHIFT
-        mov     .zz2,edx
-        movsx   ebx,.x2
-        shl     ebx,ROUND
-;if Ext>=MMX
-;        movq    mm0,.zz2M
-;;        push    .dz13    ; exchange
-;;        pop     .dz12
-;;        push    .dz23    ; exchange
-;;        pop     .dz13
-;end if
-        mov     cx,.y2
-        cmp     cx,.y3
-        jge     .ft_loop2_end
-     .ft_loop2:
-        pushad
-        push    .col
-        push    cx
-        sar     ebx,ROUND
-        push    bx
-        sar     eax,ROUND
-        push    ax          ; x1
-;if Ext>=MMX
-;        sub     esp,8
-;        movq    [esp],mm0
-;else
-        push    .zz2        ; z2 shl CATMULL_SHIFT
-        push    .zz1        ; z1 shl CATMULL_SHIFT
-;end if
-        call    flat_line_z
-        popad
-        add     eax,.dx13
-        add     ebx,.dx23
-;if  Ext>=MMX
-;        paddd   mm0,.dz23M
-;else
-        mov     edx,.dz13
-        add     .zz1,edx
-        mov     edx,.dz23
-        add     .zz2,edx
-;        mov     edx,.dz13
-;        add     .zz1,edx
-;        mov     edx,.dz12
-;        add     .zz2,edx
-;end if
-        inc     cx
-        cmp     cx,.y3
-        jl      .ft_loop2
-     .ft_loop2_end:
-        mov esp,ebp
-ret 6
-flat_line_z:
-;----------------
-;-------------in edi - pointer to screen buffer ----------------------------------
-;--------------- esi - pointer to z-buffer (each Z varible dword)-----------------
-;----------stack - (each z coordinate shifted shl CATMULL_SHIFT)------------------
-.z1 equ dword [ebp+4]
-.z2 equ dword [ebp+8]
-.x1 equ word  [ebp+12]
-.x2 equ word  [ebp+14]
-.y  equ word  [ebp+16]
-.col equ dword [ebp+18]
-.dz equ dword [ebp-4]
-     mov        ebp,esp
-;;     sub        esp,4
-     mov        ax,.y
-     or         ax,ax
-     jl         .fl_quit
-     mov        bx,[size_y_var]
-     dec        bx
-     cmp        ax,bx ;[size_y_var]
-  ;   cmp        ax,SIZE_Y-1
-     jg         .fl_quit
- ;    cmp        .x1,0
- ;    jge        .fl_ok1
- ;    cmp        .x2,0
- ;    jl         .fl_quit
- ;  .fl_ok1:
- ;    cmp        .x1,SIZE_X
- ;    jle        .fl_ok2
- ;    cmp        .x2,SIZE_X
- ;    jg         .fl_quit
- ;  .fl_ok2:
-     mov        ax,.x1
-     cmp        ax,.x2
-     je         .fl_quit
-     jl         .fl_ok
-     xchg       ax,.x2
-     mov        .x1,ax
-     mov        edx,.z1
-     xchg       edx,.z2
-     mov        .z1,edx
-   .fl_ok:
-     mov        bx,[size_x_var]
-     dec        bx
-     cmp        .x1,bx ;SIZE_X-1
-     jg         .fl_quit
-     cmp        .x2,0
-     jle        .fl_quit
-     mov        eax,.z2
-     sub        eax,.z1
-     cdq
-     mov        bx,.x2
-     sub        bx,.x1
-     movsx      ebx,bx
-     idiv       ebx
-;;     mov        .dz,eax                ; calculated delta - shifted .dz
-     push       eax
-     cmp        .x1,0
-     jge        @f
-     movsx      ebx,.x1
-     neg        ebx
-     imul       ebx
-     add        .z1,eax
-     mov        .x1,0
-    @@:
-     movzx      edx,word[size_x_var]
-     cmp        .x2,dx ;[size_x_var] ;SIZE_X
-     jl         @f
-     mov        .x2,dx ;[size_x_var] ;SIZE_X
-    @@:
- ;    movzx      edx,[size_x_var] ;SIZE_X
-     movsx      eax,.y
-     mul        edx                  ; edi = edi + (SIZE_X * y + x1)*3
-     movsx      edx,.x1
-     add        eax,edx
-     push       eax
-     lea        eax,[eax*3]
-     add        edi,eax              ; esi = esi + (SIZE_X * y + x1)*4
-     pop        eax
-     shl        eax,2
-     add        esi,eax
-     mov        cx,.x2
-     sub        cx,.x1
-     movzx      ecx,cx
-     mov        eax,.col
-     mov        ebx,.z1  ; ebx : curr. z
-     mov        edx,.dz
-     dec        ecx
-     jecxz      .draw_last
-   .ddraw:
-     cmp        ebx,dword[esi]
-   ;  cmovl      [edi],eax
-    ; cmovl      [esi],ebx
-     jge        @f
-     stosd
-     dec        edi
-     mov        dword[esi],ebx
-     jmp        .no_skip
-   @@:
-     add        edi,3
-   .no_skip:
-     add        esi,4
-     add        ebx,edx
-     loop       .ddraw
-   .draw_last:
-     cmp        ebx,dword[esi]
-     jge        .fl_quit
-     stosw
-     shr        eax,16
-     stosb
-     mov        dword[esi],ebx
-   .fl_quit:
-     mov   esp,ebp
-ret 18
+CATMULL_SHIFT equ 16
+flat_triangle_z:
+;  procedure drawing triangle with Z cordinate interpolation ------
+;  (Catmull  alghoritm)--------------------------------------------
+;  ----------------in - eax - x1 shl 16 + y1 ----------------------
+;  -------------------- ebx - x2 shl 16 + y2 ----------------------
+;  -------------------- ecx - x3 shl 16 + y3 ----------------------
+;  -------------------- edx - color 0x00RRGGBB --------------------
+;  -------------------- esi - pointer to Z-buffer -----------------
+;  -------------------- edi - pointer to screen buffer-------------
+;  -------------------- stack : z coordinates
+;  -------------------- Z-buffer : each z variable as dword
+;  -------------------- (Z coor. as word) shl CATMULL_SHIFT
+.z1     equ word[ebp+4]
+.z2     equ word[ebp+6]    ; each z coordinate as word integer
+.z3     equ word[ebp+8]
+.col    equ dword[ebp-4]
+.x1     equ word[ebp-6]
+.y1     equ word[ebp-8]
+.x2     equ word[ebp-10]
+.y2     equ word[ebp-12]
+.x3     equ word[ebp-14]
+.y3     equ word[ebp-16]
+.dx12   equ dword[ebp-20]
+;.dz12   equ dword[ebp-24]
+.dx13   equ dword[ebp-24]
+.dz13   equ dword[ebp-28]
+.dz12   equ dword[ebp-32]
+;.dz13   equ dword[ebp-32]
+.dx23   equ dword[ebp-36]
+.dz13M  equ      [ebp-40]
+.dz23   equ dword[ebp-44]
+.zz1    equ dword[ebp-48]
+.zz2    equ dword[ebp-52]
+.zz2M   equ qword[ebp-52]
+.dz12M  equ qword[ebp-32]
+.dz23M  equ qword[ebp-44]
+;if Ext>=MMX
+;        emms
+;end if
+        mov     ebp,esp
+        push    edx            ; store edx in variable .col
+      .sort2:
+        cmp     ax,bx
+        jle     .sort1
+        xchg    eax,ebx
+        mov     dx,.z1
+        xchg    dx,.z2
+        mov     .z1,dx
+      .sort1:
+        cmp     bx,cx
+        jle     .sort3
+        xchg    ebx,ecx
+        mov     dx,.z2
+        xchg    dx,.z3
+        mov     .z2,dx
+        jmp     .sort2
+      .sort3:
+        push    eax            ; store triangle coordinates in user friendly variables
+        push    ebx
+        push    ecx
+         mov      edx,80008000h  ; eax,ebx,ecx are ANDd together into edx which means that
+         and      edx,ebx        ; if *all* of them are negative a sign flag is raised
+         and      edx,ecx
+         and      edx,eax
+         test     edx,80008000h  ; Check both X&Y at once
+         jne      .ft_loop2_end
+       ;  cmp      ax,SIZE_Y
+       ;  jle      @f
+       ;  cmp      bx,SIZE_Y
+       ;  jle      @f
+       ;  cmp      cx,SIZE_Y
+       ;  jge      @f
+       ;  ror      eax,16
+       ;  ror      ebx,16
+       ;  ror      ecx,16
+       ;  cmp      ax,SIZE_X
+       ;  jle      @f
+       ;  cmp      bx,SIZE_X
+       ;  jle      @f
+       ;  cmp      cx,SIZE_X
+       ;  jle      @f
+       ;  jmp      .ft_loop2_end
+       ;@@:
+        sub     esp,52-12
+        mov     bx,.y2       ; calc delta 12
+        sub     bx,.y1
+        jnz     .ft_dx12_make
+        mov     .dx12,0
+        mov     .dz12,0
+        jmp     .ft_dx12_done
+   .ft_dx12_make:
+        mov     ax,.x2
+        sub     ax,.x1
+        cwde
+        movsx   ebx,bx
+        shl     eax,ROUND
+        cdq
+        idiv    ebx
+        mov     .dx12,eax
+        mov     ax,.z2
+        sub     ax,.z1
+        cwde
+        shl     eax,CATMULL_SHIFT
+        cdq
+        idiv    ebx
+        mov     .dz12,eax
+  .ft_dx12_done:
+        mov     bx,.y3       ; calc delta 13
+        sub     bx,.y1
+        jnz     .ft_dx13_make
+        mov     .dx13,0
+        mov     .dz13,0
+        mov     dword .dz13M,0
+        jmp     .ft_dx13_done
+   .ft_dx13_make:
+        mov     ax,.x3
+        sub     ax,.x1
+        cwde
+        movsx   ebx,bx
+        shl     eax,ROUND
+        cdq
+        idiv    ebx
+        mov     .dx13,eax
+        mov     ax,.z3
+        sub     ax,.z1
+        cwde
+        shl     eax,CATMULL_SHIFT
+        cdq
+        idiv    ebx
+        mov     .dz13,eax
+        mov     dword .dz13M,eax
+    .ft_dx13_done:
+        mov     bx,.y3       ; calc delta 23
+        sub     bx,.y2
+        jnz     .gt_dx23_make
+        mov     .dx23,0
+        mov     .dz23,0
+        jmp     .gt_dx23_done
+   .gt_dx23_make:
+        mov     ax,.x3
+        sub     ax,.x2
+        cwde
+        movsx   ebx,bx
+        shl     eax,ROUND
+        cdq
+        idiv    ebx
+        mov     .dx23,eax
+        mov     ax,.z3
+        sub     ax,.z2
+        cwde
+        shl     eax,CATMULL_SHIFT
+        cdq
+        idiv    ebx
+        mov     .dz23,eax
+    .gt_dx23_done:
+        movsx   edx,.z1
+        shl     edx,CATMULL_SHIFT
+        mov     .zz1,edx
+        mov     .zz2,edx
+        movsx   eax,.x1
+        shl     eax,ROUND    ; eax - x1
+        mov     ebx,eax      ; ebx - x2
+;if Ext>=MMX
+;        movq    mm0,.zz2M
+;end if
+        mov     cx,.y1
+        cmp     cx,.y2
+        jge     .ft_loop1_end
+     .ft_loop1:
+        pushad
+        push    .col
+        push    cx          ; y
+        sar     ebx,ROUND
+        push    bx          ; x2
+        sar     eax,ROUND
+        push    ax          ; x1
+;if Ext>=MMX
+;        sub     esp,8
+;        movq    [esp],mm0
+;else
+        push    .zz2        ; z2 shl CATMULL_SHIFT
+        push    .zz1        ; z1 shl CATMULL_SHIFT
+;end if
+        call    flat_line_z
+        popad
+        add     eax,.dx13
+        add     ebx,.dx12
+;if  Ext>=MMX
+;        paddd   mm0,.dz12M
+;else
+        mov     edx,.dz13
+        add     .zz1,edx
+        mov     edx,.dz12
+        add     .zz2,edx
+;end if
+        inc     cx
+        cmp     cx,.y2
+        jl      .ft_loop1
+    .ft_loop1_end:
+        movsx   edx,.z2
+        shl     edx,CATMULL_SHIFT
+        mov     .zz2,edx
+        movsx   ebx,.x2
+        shl     ebx,ROUND
+;if Ext>=MMX
+;        movq    mm0,.zz2M
+;;        push    .dz13    ; exchange
+;;        pop     .dz12
+;;        push    .dz23    ; exchange
+;;        pop     .dz13
+;end if
+        mov     cx,.y2
+        cmp     cx,.y3
+        jge     .ft_loop2_end
+     .ft_loop2:
+        pushad
+        push    .col
+        push    cx
+        sar     ebx,ROUND
+        push    bx
+        sar     eax,ROUND
+        push    ax          ; x1
+;if Ext>=MMX
+;        sub     esp,8
+;        movq    [esp],mm0
+;else
+        push    .zz2        ; z2 shl CATMULL_SHIFT
+        push    .zz1        ; z1 shl CATMULL_SHIFT
+;end if
+        call    flat_line_z
+        popad
+        add     eax,.dx13
+        add     ebx,.dx23
+;if  Ext>=MMX
+;        paddd   mm0,.dz23M
+;else
+        mov     edx,.dz13
+        add     .zz1,edx
+        mov     edx,.dz23
+        add     .zz2,edx
+;        mov     edx,.dz13
+;        add     .zz1,edx
+;        mov     edx,.dz12
+;        add     .zz2,edx
+;end if
+        inc     cx
+        cmp     cx,.y3
+        jl      .ft_loop2
+     .ft_loop2_end:
+        mov esp,ebp
+ret 6
+flat_line_z:
+;----------------
+;-------------in edi - pointer to screen buffer ----------------------------------
+;--------------- esi - pointer to z-buffer (each Z varible dword)-----------------
+;----------stack - (each z coordinate shifted shl CATMULL_SHIFT)------------------
+.z1 equ dword [ebp+4]
+.z2 equ dword [ebp+8]
+.x1 equ word  [ebp+12]
+.x2 equ word  [ebp+14]
+.y  equ word  [ebp+16]
+.col equ dword [ebp+18]
+.dz equ dword [ebp-4]
+     mov        ebp,esp
+;;     sub        esp,4
+     mov        ax,.y
+     or         ax,ax
+     jl         .fl_quit
+     mov        bx,[size_y_var]
+     dec        bx
+     cmp        ax,bx ;[size_y_var]
+  ;   cmp        ax,SIZE_Y-1
+     jg         .fl_quit
+ ;    cmp        .x1,0
+ ;    jge        .fl_ok1
+ ;    cmp        .x2,0
+ ;    jl         .fl_quit
+ ;  .fl_ok1:
+ ;    cmp        .x1,SIZE_X
+ ;    jle        .fl_ok2
+ ;    cmp        .x2,SIZE_X
+ ;    jg         .fl_quit
+ ;  .fl_ok2:
+     mov        ax,.x1
+     cmp        ax,.x2
+     je         .fl_quit
+     jl         .fl_ok
+     xchg       ax,.x2
+     mov        .x1,ax
+     mov        edx,.z1
+     xchg       edx,.z2
+     mov        .z1,edx
+   .fl_ok:
+     mov        bx,[size_x_var]
+     dec        bx
+     cmp        .x1,bx ;SIZE_X-1
+     jg         .fl_quit
+     cmp        .x2,0
+     jle        .fl_quit
+     mov        eax,.z2
+     sub        eax,.z1
+     cdq
+     mov        bx,.x2
+     sub        bx,.x1
+     movsx      ebx,bx
+     idiv       ebx
+;;     mov        .dz,eax                ; calculated delta - shifted .dz
+     push       eax
+     cmp        .x1,0
+     jge        @f
+     movsx      ebx,.x1
+     neg        ebx
+     imul       ebx
+     add        .z1,eax
+     mov        .x1,0
+    @@:
+     movzx      edx,word[size_x_var]
+     cmp        .x2,dx ;[size_x_var] ;SIZE_X
+     jl         @f
+     mov        .x2,dx ;[size_x_var] ;SIZE_X
+    @@:
+ ;    movzx      edx,[size_x_var] ;SIZE_X
+     movsx      eax,.y
+     mul        edx                  ; edi = edi + (SIZE_X * y + x1)*3
+     movsx      edx,.x1
+     add        eax,edx
+     push       eax
+     lea        eax,[eax*3]
+     add        edi,eax              ; esi = esi + (SIZE_X * y + x1)*4
+     pop        eax
+     shl        eax,2
+     add        esi,eax
+     mov        cx,.x2
+     sub        cx,.x1
+     movzx      ecx,cx
+     mov        eax,.col
+     mov        ebx,.z1  ; ebx : curr. z
+     mov        edx,.dz
+     dec        ecx
+     jecxz      .draw_last
+   .ddraw:
+     cmp        ebx,dword[esi]
+   ;  cmovl      [edi],eax
+    ; cmovl      [esi],ebx
+     jge        @f
+     stosd
+     dec        edi
+     mov        dword[esi],ebx
+     jmp        .no_skip
+   @@:
+     add        edi,3
+   .no_skip:
+     add        esi,4
+     add        ebx,edx
+     loop       .ddraw
+   .draw_last:
+     cmp        ebx,dword[esi]
+     jge        .fl_quit
+     stosw
+     shr        eax,16
+     stosb
+     mov        dword[esi],ebx
+   .fl_quit:
+     mov   esp,ebp
+ret 18

 /programs/demos/view3ds/grd_cat.inc
 ,704 → 1,704
-ROUND equ 8
-CATMULL_SHIFT equ 8
-gouraud_triangle_z:
-;----procedure drawing gouraud triangle with z coordinate
-;----interpolation ( Catmull alghoritm )-----------------
-;------------------in - eax - x1 shl 16 + y1 ------------
-;---------------------- ebx - x2 shl 16 + y2 ------------
-;---------------------- ecx - x3 shl 16 + y3 ------------
-;---------------------- esi - pointer to Z-buffer--------
-;---------------------- Z-buffer filled with dd variables
-;---------------------- shifted CATMULL_SHIFT------------
-;---------------------- edi - pointer to screen buffer---
-;---------------------- stack : colors-------------------
-;----------------- procedure don't save registers !!-----
-.col1r equ ebp+4   ; each color as word
-.col1g equ ebp+6   ; each z coordinate as word
-.col1b equ ebp+8
-.z1    equ ebp+10
-.col2r equ ebp+12
-.col2g equ ebp+14
-.col2b equ ebp+16
-.z2    equ ebp+18
-.col3r equ ebp+20
-.col3g equ ebp+22
-.col3b equ ebp+24
-.z3    equ ebp+26
-.x1    equ word[ebp-2]
-.y1    equ word[ebp-4]
-.x2    equ word[ebp-6]
-.y2    equ word[ebp-8]
-.x3    equ word[ebp-10]
-.y3    equ word[ebp-12]
-.dx12  equ dword[ebp-16]
-.dz12  equ dword[ebp-20]
-.dc12r equ dword[ebp-24]
-.dc12g equ dword[ebp-28]
-.dc12b equ dword[ebp-32]
-.dx13  equ dword[ebp-36]
-.dz13  equ dword[ebp-40]
-.dc13r equ dword[ebp-44]
-.dc13g equ dword[ebp-48]
-.dc13b equ dword[ebp-52]
-.dx23  equ dword[ebp-56]
-.dz23  equ dword[ebp-60]
-.dc23r equ dword[ebp-64]
-.dc23g equ dword[ebp-68]
-.dc23b equ dword[ebp-72]
-.zz1   equ dword[ebp-76]
-.c1r   equ dword[ebp-80]
-.c1g   equ dword[ebp-84]
-.c1b   equ dword[ebp-88]
-.zz2   equ dword[ebp-92]
-.c2r   equ dword[ebp-96]
-.c2g   equ dword[ebp-100]
-.c2b   equ dword[ebp-104]
-;.zz1   equ dword[ebp-100]
-;.zz2   equ dword[ebp-104]
-.c1bM equ [ebp-88]
-.c2bM equ [ebp-104]
-.c1rM equ [ebp-80]
-.c2rM equ [ebp-96]
-.dc23bM equ [ebp-72]
-.dc13bM equ [ebp-52]
-.dc12bM equ [ebp-32]
-.dc12rM equ [ebp-24]
-.dc13rM equ [ebp-44]
-.dc23rM equ [ebp-64]
-if Ext=MMX
-      emms
-end if
-       mov     ebp,esp
-     ;  sub     esp,84
- .sort3:                  ; sort triangle coordinates...
-       cmp     ax,bx
-       jle     .sort1
-       xchg    eax,ebx
-       mov     edx,dword[.col1r]
-       xchg    edx,dword[.col2r]
-       mov     dword[.col1r],edx
-       mov     edx,dword[.col1b]
-       xchg    edx,dword[.col2b]
-       mov     dword[.col1b],edx
- .sort1:
-       cmp      bx,cx
-       jle      .sort2
-       xchg     ebx,ecx
-       mov      edx,dword[.col2r]
-       xchg     edx,dword[.col3r]
-       mov      dword[.col2r],edx
-       mov      edx,dword[.col2b]
-       xchg     edx,dword[.col3b]
-       mov      dword[.col2b],edx
-       jmp .sort3
- .sort2:
-       push     eax          ; store in variables
-       push     ebx
-       push     ecx
-         mov      edx,80008000h  ; eax,ebx,ecx are ANDd together into edx which means that
-         and      edx,ebx        ; if *all* of them are negative a sign flag is raised
-         and      edx,ecx
-         and      edx,eax
-         test     edx,80008000h  ; Check both X&Y at once
-         jne      .gt_loop2_end
-       mov      bx,.y2       ; calc deltas
-       sub      bx,.y1
-       jnz      .gt_dx12_make
-      ; mov      .dx12,0
-      ; mov      .dz12,0
-      ; mov      .dc12r,0
-      ; mov      .dc12g,0
-      ; mov      .dc12b,0
-       mov      ecx,5
-     @@:
-       push     dword 0
-       loop     @b
-       jmp      .gt_dx12_done
-  .gt_dx12_make:
-       mov      ax,.x2
-       sub      ax,.x1
-       cwde
-       movsx    ebx,bx
-       shl      eax,ROUND
-       cdq
-       idiv     ebx
- ;      mov      .dx12,eax
-       push      eax
-       mov      ax,word[.z2]
-       sub      ax,word[.z1]
-       cwde
-       shl      eax,CATMULL_SHIFT
-       cdq
-       idiv     ebx
-       push     eax
-       mov      ax,word[.col2r]
-       sub      ax,word[.col1r]
-       cwde
-       shl      eax,ROUND
-       cdq
-       idiv     ebx
-      ; mov      .dc12r,eax
-       push       eax
-       mov        ax,word[.col2g]
-       sub        ax,word[.col1g]
-       cwde
-       shl      eax,ROUND
-       cdq
-       idiv     ebx
-     ;  mov .dc12g,eax
-       push     eax
-       mov      ax,word[.col2b]        ;;---
-       sub      ax,word[.col1b]
-       cwde
-       shl      eax,ROUND
-       cdq
-       idiv     ebx
-      ; mov .dc12b,eax
-       push     eax
-   .gt_dx12_done:
-       mov      bx,.y3       ; calc deltas
-       sub      bx,.y1
-       jnz      .gt_dx13_make
-      ; mov      .dx13,0
-      ; mov      .dz13,0
-      ; mov      .dc13r,0
-      ; mov      .dc13g,0
-      ; mov      .dc13b,0
-       mov      ecx,5
-     @@:
-       push     dword 0
-       loop     @b
-       jmp      .gt_dx13_done
-    .gt_dx13_make:
-       mov      ax,.x3
-       sub      ax,.x1
-       cwde
-       movsx    ebx,bx
-       shl      eax,ROUND
-       cdq
-       idiv     ebx
- ;      mov      .dx13,eax
-       push      eax
-       mov      ax,word[.z3]
-       sub      ax,word[.z1]
-       cwde
-       shl      eax,CATMULL_SHIFT
-       cdq
-       idiv     ebx
-       push     eax
-       mov      ax,word[.col3r]
-       sub      ax,word[.col1r]
-       cwde
-       shl      eax,ROUND
-       cdq
-       idiv     ebx
-      ; mov      .dc13r,eax
-       push       eax
-       mov        ax,word[.col3g]
-       sub        ax,word[.col1g]
-       cwde
-       shl      eax,ROUND
-       cdq
-       idiv     ebx
-     ;  mov .dc13g,eax
-       push     eax
-       mov      ax,word[.col3b]
-       sub      ax,word[.col1b]
-       cwde
-       shl      eax,ROUND
-       cdq
-       idiv     ebx
-      ; mov .dc13b,eax
-       push     eax
-   .gt_dx13_done:
-       mov      bx,.y3       ; calc deltas
-       sub      bx,.y2
-       jnz      .gt_dx23_make
-      ; mov      .dx23,0
-      ; mov      .dz23,0
-      ; mov      .dc23r,0
-      ; mov      .dc23g,0
-      ; mov      .dc23b,0
-       mov      ecx,5
-     @@:
-       push     dword 0
-       loop     @b
-       jmp      .gt_dx23_done
-    .gt_dx23_make:
-       mov      ax,.x3
-       sub      ax,.x2
-       cwde
-       movsx    ebx,bx
-       shl      eax,ROUND
-       cdq
-       idiv     ebx
- ;      mov      .dx23,eax
-       push      eax
-       mov      ax,word[.z3]
-       sub      ax,word[.z2]
-       cwde
-       shl      eax,CATMULL_SHIFT
-       cdq
-       idiv     ebx
-       push     eax
-       mov      ax,word[.col3r]
-       sub      ax,word[.col2r]
-       cwde
-       shl      eax,ROUND
-       cdq
-       idiv     ebx
-      ; mov     .dc23r,eax
-       push     eax
-       mov      ax,word[.col3g]
-       sub      ax,word[.col2g]
-       cwde
-       shl      eax,ROUND
-       cdq
-       idiv     ebx
-     ;  mov .dc23g,eax
-       push     eax
-       mov      ax,word[.col3b]
-       sub      ax,word[.col2b]
-       cwde
-       shl      eax,ROUND
-       cdq
-       idiv     ebx
-      ; mov .dc23b,eax
-       push     eax
-   .gt_dx23_done:
-       sub      esp,32
-       movsx    eax,.x1                    ; eax - cur x1
-       shl      eax,ROUND                  ; ebx - cur x2
-       mov      ebx,eax
-       movsx    edx,word[.z1]
-       shl      edx,CATMULL_SHIFT
-       mov      .zz1,edx
-       mov      .zz2,edx
-       movzx    edx,word[.col1r]
-       shl      edx,ROUND
-       mov      .c1r,edx
-       mov      .c2r,edx
-       movzx    edx,word[.col1g]
-       shl      edx,ROUND
-       mov      .c1g,edx
-       mov      .c2g,edx
-       movzx    edx,word[.col1b]
-       shl      edx,ROUND
-       mov      .c1b,edx
-       mov      .c2b,edx
-       mov      cx,.y1
-       cmp      cx,.y2
-       jge      .gt_loop1_end
-    .gt_loop1:
-       pushad
-    ; macro .debug
-       mov      edx,.c2r              ; c2r,c2g,c2b,c1r,c1g,c1b - current colors
-       sar      edx,ROUND
-       push     dx
-       mov      edx,.c2g
-       sar      edx,ROUND
-       push     dx
-       mov      edx,.c2b
-       sar      edx,ROUND
-       push     dx
-       sar      ebx,ROUND    ; x2
-       push     bx
-       mov      edx,.c1r
-       sar      edx,ROUND
-       push     dx
-       mov      edx,.c1g
-       sar      edx,ROUND
-       push     dx
-       mov      edx,.c1b
-       sar      edx,ROUND
-       push     dx
-       sar      eax,ROUND
-       push     ax            ; x1
-       push     cx            ; y
-       push     .zz2
-       push     .zz1
-       call     gouraud_line_z
-       popad
-if Ext >= MMX
-       movq     mm0,.c1bM
-       paddd    mm0,qword .dc13bM
-       movq     .c1bM,mm0
-       movq     mm1,.c2bM
-       paddd    mm1,qword .dc12bM
-       movq     .c2bM,mm1
-       movq     mm0,.c1rM
-       paddd    mm0,qword .dc13rM
-       movq     .c1rM,mm0
-       movq     mm1,.c2rM
-       paddd    mm1,qword .dc12rM
-       movq     .c2rM,mm1
-else
-       mov      edx,.dc13r
-       add      .c1r,edx
-       mov      edx,.dc13g
-       add      .c1g,edx
-       mov      edx,.dc13b
-       add      .c1b,edx
-       mov      edx,.dc12r
-       add      .c2r,edx
-       mov      edx,.dc12g
-       add      .c2g,edx
-       mov      edx,.dc12b
-       add      .c2b,edx
-       mov      edx,.dz13
-       add      .zz1,edx
-       mov      edx,.dz12
-       add      .zz2,edx
-end if
-       add      eax,.dx13
-       add      ebx,.dx12
-       inc      cx
-       cmp      cx,.y2
-       jl       .gt_loop1
-   .gt_loop1_end:
-       mov      cx,.y2
-       cmp      cx,.y3
-       jge      .gt_loop2_end
-       movsx    ebx,.x2                    ; eax - cur x1
-       shl      ebx,ROUND                  ; ebx - cur x2
-       movsx    edx,word[.z2]
-       shl      edx,CATMULL_SHIFT
-       mov      .zz2,edx
-       movzx    edx,word[.col2r]
-       shl      edx,ROUND
-       mov      .c2r,edx
-       movzx    edx,word[.col2g]
-       shl      edx,ROUND
-       mov      .c2g,edx
-       movzx    edx,word[.col2b]
-       shl      edx,ROUND
-       mov      .c2b,edx
-    .gt_loop2:
-       pushad
-    ; macro .debug
-       mov      edx,.c2r              ; c2r,c2g,c2b,c1r,c1g,c1b - current colors
-       sar      edx,ROUND
-       push     dx
-       mov      edx,.c2g
-       sar      edx,ROUND
-       push     dx
-       mov      edx,.c2b
-       sar      edx,ROUND
-       push     dx
-       sar      ebx,ROUND    ; x2
-       push     bx
-       mov      edx,.c1r
-       sar      edx,ROUND
-       push     dx
-       mov      edx,.c1g
-       sar      edx,ROUND
-       push     dx
-       mov      edx,.c1b
-       sar      edx,ROUND
-       push     dx
-       sar      eax,ROUND
-       push     ax            ; x1
-       push     cx            ; y
-       push     .zz2
-       push     .zz1
-       call     gouraud_line_z
-       popad
-if Ext >= MMX
-       movq     mm0,.c1bM
-       paddd    mm0,qword .dc13bM
-       movq     .c1bM,mm0
-       movq     mm1,.c2bM
-       paddd    mm1,qword .dc23bM
-       movq     .c2bM,mm1
-       movq     mm0,.c1rM
-       paddd    mm0,qword .dc13rM
-       movq     .c1rM,mm0
-       movq     mm1,.c2rM
-       paddd    mm1,qword .dc23rM
-       movq     .c2rM,mm1
-else
-       mov      edx,.dc13r
-       add      .c1r,edx
-       mov      edx,.dc13g
-       add      .c1g,edx
-       mov      edx,.dc13b
-       add      .c1b,edx
-       mov      edx,.dc23r
-       add      .c2r,edx
-       mov      edx,.dc23g
-       add      .c2g,edx
-       mov      edx,.dc23b
-       add      .c2b,edx
-       mov      edx,.dz13
-       add      .zz1,edx
-       mov      edx,.dz23
-       add      .zz2,edx
-end if
-       add      eax,.dx13
-       add      ebx,.dx23
-       inc      cx
-       cmp      cx,.y3
-       jl       .gt_loop2
-   .gt_loop2_end:
-       mov      esp,ebp
-ret 24
-gouraud_line_z:
-;----------------- procedure drawing gouraud line
-;----------------- with z coordinate interpolation
-;----------------- esi - pointer to Z_buffer
-;----------------- edi - pointer to screen buffer
-;----------------- stack:
-.z1  equ dword[ebp+4]   ; z coordiunate shifted left CATMULL_SHIFT
-.z2  equ dword[ebp+8]
-.y   equ word[ebp+12]
-.x1  equ ebp+14
-.c1b equ ebp+16
-.c1g equ ebp+18
-.c1r equ ebp+20
-.x2  equ ebp+22
-.c2b equ ebp+24
-.c2g equ ebp+26
-.c2r equ ebp+28
-.dz   equ dword[ebp-4]
-.dc_b equ dword[ebp-8]
-.dc_g equ dword[ebp-12]
-.dc_r equ dword[ebp-16]
-.c_z  equ dword[ebp-20]
-.cb   equ dword[ebp-24]
-.cg   equ dword[ebp-28]
-.cr   equ dword[ebp-32]
-;.cg2  equ dword[ebp-36]
-.crM  equ ebp-32
-.cgM  equ ebp-28
-.cbM  equ ebp-24
-.dc_rM equ ebp-16
-.dc_gM equ ebp-12
-.dc_bM equ ebp-8
-        mov       ebp,esp
-        mov     ax,.y
-        or      ax,ax
-        jl      .gl_quit
-        mov     bx,[size_y_var]
-        dec     bx
-        cmp     ax,bx ;SIZE_Y
-        jge     .gl_quit
-        mov     eax,dword[.x1]
-        cmp     ax,word[.x2]
-        je      .gl_quit
-        jl      @f
-        xchg    eax,dword[.x2]
-        mov     dword[.x1],eax
-        mov     eax,dword[.c1g]
-        xchg    eax,dword[.c2g]
-        mov     dword[.c1g],eax
-        mov     eax,.z1
-        xchg    eax,.z2
-        mov     .z1,eax
-   @@:
-        mov     bx,[size_x_var]
-        dec     bx
-        cmp     word[.x1],bx  ;SIZE_X
-        jge     .gl_quit
-        cmp     word[.x2],0
-        jle     .gl_quit
-        mov     eax,.z2
-        sub     eax,.z1
-        cdq
-        mov     bx,word[.x2]      ; dz = z2-z1/x2-x1
-        sub     bx,word[.x1]
-        movsx   ebx,bx
-        idiv    ebx
-        push    eax
-        mov     ax,word[.c2b]
-        sub     ax,word[.c1b]
-        cwde
-        shl     eax,ROUND
-        cdq
-        idiv    ebx
-        push    eax
-        mov     ax,word[.c2g]
-        sub     ax,word[.c1g]
-        cwde
-        shl     eax,ROUND
-        cdq
-        idiv    ebx
-        push    eax
-        mov     ax,word[.c2r]
-        sub     ax,word[.c1r]
-        cwde
-        shl     eax,ROUND         ; dc_r = c2r-c1r/x2-x1
-        cdq
-        idiv    ebx
-        push    eax
-        cmp     word[.x1],0     ; clipping on function
-        jg      @f
-        mov     eax,.dz
-        movsx   ebx,word[.x1]
-        neg     ebx
-        imul    ebx
-        add     .z1,eax
-        mov     word[.x1],0
-        mov     eax,.dc_r
-        imul    ebx
-        sar     eax,ROUND
-        add     word[.c1r],ax
-        mov     eax,.dc_g
-        imul    ebx
-        sar     eax,ROUND
-        add     word[.c1g],ax
-        mov     eax,.dc_b
-        imul    ebx
-        sar     eax,ROUND
-        add     word[.c1b],ax
-      @@:
-        mov     bx,[size_x_var]
-        dec     bx
-        cmp     word[.x2],bx  ;SIZE_X
-        jl      @f
-        mov     word[.x2],bx  ;SIZE_X
-     @@:
-        sub     esp,16      ; calculate memory begin
-        movzx   edx,word[size_x_var]  ;SIZE_X       ; in buffers
-        movzx   eax,.y
-        mul     edx
-        movzx   edx,word[.x1]
-        add     eax,edx
-        push    eax
-        lea     eax,[eax*3]
-        add     edi,eax
-        pop     eax
-        shl     eax,2
-        add     esi,eax
-        mov     cx,word[.x2]
-        sub     cx,word[.x1]
-        movzx   ecx,cx
-        mov     ebx,.z1          ; ebx - currrent z shl CATMULL_SIFT
-;if Ext >= SSE
-;        mov     .cz,edx
-;end if
-        mov     edx,.dz          ; edx - delta z
-        movzx   eax,word[.c1r]
-        shl     eax,ROUND
-        mov     .cr,eax
-        movzx   eax,word[.c1g]
-        shl     eax,ROUND
-        mov     .cg,eax
-        movzx   eax,word[.c1b]
-        shl     eax,ROUND
-        mov     .cb,eax
-if Ext = MMX
-;        mov     .c_z,edx
-        movd    mm2,[.dc_bM]         ; delta color blue MMX
-        movd    mm3,[.cbM]           ; current blue MMX
-        movq    mm5,[.dc_rM]
-        movq    mm4,[.crM]
-        pxor    mm6,mm6
-end if
-      .ddraw:
-;if Ext = MMX
-;        movq    mm0,mm3
-;        psrsq   mm0,32
-;        movd    ebx,mm0
-;end if
-        cmp     ebx,dword[esi]   ; esi - z_buffer
-        jge     @f               ; edi - Screen buffer
-if Ext = MMX
-        movq    mm0,mm3          ; mm0, mm1 - temp registers
-        psrld   mm0,ROUND
-        movq    mm1,mm4
-        psrld   mm1,ROUND
-        packssdw  mm1,mm0
-        packuswb  mm1,mm6
-;        movd     [edi],mm1
-        movd      eax,mm1
-        stosw
-        shr       eax,16
-        stosb
-else
-        mov     eax,.cr
-        sar     eax,ROUND
-        stosb
-        mov     eax,.cg
-        sar     eax,ROUND
-        stosb
-        mov     eax,.cb
-        sar     eax,ROUND
-        stosb
-end if
-        mov     dword[esi],ebx
-;if Ext = NON
-        jmp     .no_skip
-;end if
-      @@:
-        add     edi,3
-      .no_skip:
-        add     esi,4
-;if Ext=NON
-        add     ebx,edx
-;end if
-if Ext=MMX
-        paddd   mm3,mm2
-        paddd   mm4,mm5
-else
-        mov     eax,.dc_g
-        add     .cg,eax
-        mov     eax,.dc_b
-        add     .cb,eax
-        mov     eax,.dc_r
-        add     .cr,eax
-end if
-        loop    .ddraw
-   .gl_quit:
-        mov       esp,ebp
-ret 26
+ROUND equ 8
+CATMULL_SHIFT equ 8
+gouraud_triangle_z:
+;----procedure drawing gouraud triangle with z coordinate
+;----interpolation ( Catmull alghoritm )-----------------
+;------------------in - eax - x1 shl 16 + y1 ------------
+;---------------------- ebx - x2 shl 16 + y2 ------------
+;---------------------- ecx - x3 shl 16 + y3 ------------
+;---------------------- esi - pointer to Z-buffer--------
+;---------------------- Z-buffer filled with dd variables
+;---------------------- shifted CATMULL_SHIFT------------
+;---------------------- edi - pointer to screen buffer---
+;---------------------- stack : colors-------------------
+;----------------- procedure don't save registers !!-----
+.col1r equ ebp+4   ; each color as word
+.col1g equ ebp+6   ; each z coordinate as word
+.col1b equ ebp+8
+.z1    equ ebp+10
+.col2r equ ebp+12
+.col2g equ ebp+14
+.col2b equ ebp+16
+.z2    equ ebp+18
+.col3r equ ebp+20
+.col3g equ ebp+22
+.col3b equ ebp+24
+.z3    equ ebp+26
+.x1    equ word[ebp-2]
+.y1    equ word[ebp-4]
+.x2    equ word[ebp-6]
+.y2    equ word[ebp-8]
+.x3    equ word[ebp-10]
+.y3    equ word[ebp-12]
+.dx12  equ dword[ebp-16]
+.dz12  equ dword[ebp-20]
+.dc12r equ dword[ebp-24]
+.dc12g equ dword[ebp-28]
+.dc12b equ dword[ebp-32]
+.dx13  equ dword[ebp-36]
+.dz13  equ dword[ebp-40]
+.dc13r equ dword[ebp-44]
+.dc13g equ dword[ebp-48]
+.dc13b equ dword[ebp-52]
+.dx23  equ dword[ebp-56]
+.dz23  equ dword[ebp-60]
+.dc23r equ dword[ebp-64]
+.dc23g equ dword[ebp-68]
+.dc23b equ dword[ebp-72]
+.zz1   equ dword[ebp-76]
+.c1r   equ dword[ebp-80]
+.c1g   equ dword[ebp-84]
+.c1b   equ dword[ebp-88]
+.zz2   equ dword[ebp-92]
+.c2r   equ dword[ebp-96]
+.c2g   equ dword[ebp-100]
+.c2b   equ dword[ebp-104]
+;.zz1   equ dword[ebp-100]
+;.zz2   equ dword[ebp-104]
+.c1bM equ [ebp-88]
+.c2bM equ [ebp-104]
+.c1rM equ [ebp-80]
+.c2rM equ [ebp-96]
+.dc23bM equ [ebp-72]
+.dc13bM equ [ebp-52]
+.dc12bM equ [ebp-32]
+.dc12rM equ [ebp-24]
+.dc13rM equ [ebp-44]
+.dc23rM equ [ebp-64]
+if Ext=MMX
+      emms
+end if
+       mov     ebp,esp
+     ;  sub     esp,84
+ .sort3:                  ; sort triangle coordinates...
+       cmp     ax,bx
+       jle     .sort1
+       xchg    eax,ebx
+       mov     edx,dword[.col1r]
+       xchg    edx,dword[.col2r]
+       mov     dword[.col1r],edx
+       mov     edx,dword[.col1b]
+       xchg    edx,dword[.col2b]
+       mov     dword[.col1b],edx
+ .sort1:
+       cmp      bx,cx
+       jle      .sort2
+       xchg     ebx,ecx
+       mov      edx,dword[.col2r]
+       xchg     edx,dword[.col3r]
+       mov      dword[.col2r],edx
+       mov      edx,dword[.col2b]
+       xchg     edx,dword[.col3b]
+       mov      dword[.col2b],edx
+       jmp .sort3
+ .sort2:
+       push     eax          ; store in variables
+       push     ebx
+       push     ecx
+         mov      edx,80008000h  ; eax,ebx,ecx are ANDd together into edx which means that
+         and      edx,ebx        ; if *all* of them are negative a sign flag is raised
+         and      edx,ecx
+         and      edx,eax
+         test     edx,80008000h  ; Check both X&Y at once
+         jne      .gt_loop2_end
+       mov      bx,.y2       ; calc deltas
+       sub      bx,.y1
+       jnz      .gt_dx12_make
+      ; mov      .dx12,0
+      ; mov      .dz12,0
+      ; mov      .dc12r,0
+      ; mov      .dc12g,0
+      ; mov      .dc12b,0
+       mov      ecx,5
+     @@:
+       push     dword 0
+       loop     @b
+       jmp      .gt_dx12_done
+  .gt_dx12_make:
+       mov      ax,.x2
+       sub      ax,.x1
+       cwde
+       movsx    ebx,bx
+       shl      eax,ROUND
+       cdq
+       idiv     ebx
+ ;      mov      .dx12,eax
+       push      eax
+       mov      ax,word[.z2]
+       sub      ax,word[.z1]
+       cwde
+       shl      eax,CATMULL_SHIFT
+       cdq
+       idiv     ebx
+       push     eax
+       mov      ax,word[.col2r]
+       sub      ax,word[.col1r]
+       cwde
+       shl      eax,ROUND
+       cdq
+       idiv     ebx
+      ; mov      .dc12r,eax
+       push       eax
+       mov        ax,word[.col2g]
+       sub        ax,word[.col1g]
+       cwde
+       shl      eax,ROUND
+       cdq
+       idiv     ebx
+     ;  mov .dc12g,eax
+       push     eax
+       mov      ax,word[.col2b]        ;;---
+       sub      ax,word[.col1b]
+       cwde
+       shl      eax,ROUND
+       cdq
+       idiv     ebx
+      ; mov .dc12b,eax
+       push     eax
+   .gt_dx12_done:
+       mov      bx,.y3       ; calc deltas
+       sub      bx,.y1
+       jnz      .gt_dx13_make
+      ; mov      .dx13,0
+      ; mov      .dz13,0
+      ; mov      .dc13r,0
+      ; mov      .dc13g,0
+      ; mov      .dc13b,0
+       mov      ecx,5
+     @@:
+       push     dword 0
+       loop     @b
+       jmp      .gt_dx13_done
+    .gt_dx13_make:
+       mov      ax,.x3
+       sub      ax,.x1
+       cwde
+       movsx    ebx,bx
+       shl      eax,ROUND
+       cdq
+       idiv     ebx
+ ;      mov      .dx13,eax
+       push      eax
+       mov      ax,word[.z3]
+       sub      ax,word[.z1]
+       cwde
+       shl      eax,CATMULL_SHIFT
+       cdq
+       idiv     ebx
+       push     eax
+       mov      ax,word[.col3r]
+       sub      ax,word[.col1r]
+       cwde
+       shl      eax,ROUND
+       cdq
+       idiv     ebx
+      ; mov      .dc13r,eax
+       push       eax
+       mov        ax,word[.col3g]
+       sub        ax,word[.col1g]
+       cwde
+       shl      eax,ROUND
+       cdq
+       idiv     ebx
+     ;  mov .dc13g,eax
+       push     eax
+       mov      ax,word[.col3b]
+       sub      ax,word[.col1b]
+       cwde
+       shl      eax,ROUND
+       cdq
+       idiv     ebx
+      ; mov .dc13b,eax
+       push     eax
+   .gt_dx13_done:
+       mov      bx,.y3       ; calc deltas
+       sub      bx,.y2
+       jnz      .gt_dx23_make
+      ; mov      .dx23,0
+      ; mov      .dz23,0
+      ; mov      .dc23r,0
+      ; mov      .dc23g,0
+      ; mov      .dc23b,0
+       mov      ecx,5
+     @@:
+       push     dword 0
+       loop     @b
+       jmp      .gt_dx23_done
+    .gt_dx23_make:
+       mov      ax,.x3
+       sub      ax,.x2
+       cwde
+       movsx    ebx,bx
+       shl      eax,ROUND
+       cdq
+       idiv     ebx
+ ;      mov      .dx23,eax
+       push      eax
+       mov      ax,word[.z3]
+       sub      ax,word[.z2]
+       cwde
+       shl      eax,CATMULL_SHIFT
+       cdq
+       idiv     ebx
+       push     eax
+       mov      ax,word[.col3r]
+       sub      ax,word[.col2r]
+       cwde
+       shl      eax,ROUND
+       cdq
+       idiv     ebx
+      ; mov     .dc23r,eax
+       push     eax
+       mov      ax,word[.col3g]
+       sub      ax,word[.col2g]
+       cwde
+       shl      eax,ROUND
+       cdq
+       idiv     ebx
+     ;  mov .dc23g,eax
+       push     eax
+       mov      ax,word[.col3b]
+       sub      ax,word[.col2b]
+       cwde
+       shl      eax,ROUND
+       cdq
+       idiv     ebx
+      ; mov .dc23b,eax
+       push     eax
+   .gt_dx23_done:
+       sub      esp,32
+       movsx    eax,.x1                    ; eax - cur x1
+       shl      eax,ROUND                  ; ebx - cur x2
+       mov      ebx,eax
+       movsx    edx,word[.z1]
+       shl      edx,CATMULL_SHIFT
+       mov      .zz1,edx
+       mov      .zz2,edx
+       movzx    edx,word[.col1r]
+       shl      edx,ROUND
+       mov      .c1r,edx
+       mov      .c2r,edx
+       movzx    edx,word[.col1g]
+       shl      edx,ROUND
+       mov      .c1g,edx
+       mov      .c2g,edx
+       movzx    edx,word[.col1b]
+       shl      edx,ROUND
+       mov      .c1b,edx
+       mov      .c2b,edx
+       mov      cx,.y1
+       cmp      cx,.y2
+       jge      .gt_loop1_end
+    .gt_loop1:
+       pushad
+    ; macro .debug
+       mov      edx,.c2r              ; c2r,c2g,c2b,c1r,c1g,c1b - current colors
+       sar      edx,ROUND
+       push     dx
+       mov      edx,.c2g
+       sar      edx,ROUND
+       push     dx
+       mov      edx,.c2b
+       sar      edx,ROUND
+       push     dx
+       sar      ebx,ROUND    ; x2
+       push     bx
+       mov      edx,.c1r
+       sar      edx,ROUND
+       push     dx
+       mov      edx,.c1g
+       sar      edx,ROUND
+       push     dx
+       mov      edx,.c1b
+       sar      edx,ROUND
+       push     dx
+       sar      eax,ROUND
+       push     ax            ; x1
+       push     cx            ; y
+       push     .zz2
+       push     .zz1
+       call     gouraud_line_z
+       popad
+if Ext >= MMX
+       movq     mm0,.c1bM
+       paddd    mm0,qword .dc13bM
+       movq     .c1bM,mm0
+       movq     mm1,.c2bM
+       paddd    mm1,qword .dc12bM
+       movq     .c2bM,mm1
+       movq     mm0,.c1rM
+       paddd    mm0,qword .dc13rM
+       movq     .c1rM,mm0
+       movq     mm1,.c2rM
+       paddd    mm1,qword .dc12rM
+       movq     .c2rM,mm1
+else
+       mov      edx,.dc13r
+       add      .c1r,edx
+       mov      edx,.dc13g
+       add      .c1g,edx
+       mov      edx,.dc13b
+       add      .c1b,edx
+       mov      edx,.dc12r
+       add      .c2r,edx
+       mov      edx,.dc12g
+       add      .c2g,edx
+       mov      edx,.dc12b
+       add      .c2b,edx
+       mov      edx,.dz13
+       add      .zz1,edx
+       mov      edx,.dz12
+       add      .zz2,edx
+end if
+       add      eax,.dx13
+       add      ebx,.dx12
+       inc      cx
+       cmp      cx,.y2
+       jl       .gt_loop1
+   .gt_loop1_end:
+       mov      cx,.y2
+       cmp      cx,.y3
+       jge      .gt_loop2_end
+       movsx    ebx,.x2                    ; eax - cur x1
+       shl      ebx,ROUND                  ; ebx - cur x2
+       movsx    edx,word[.z2]
+       shl      edx,CATMULL_SHIFT
+       mov      .zz2,edx
+       movzx    edx,word[.col2r]
+       shl      edx,ROUND
+       mov      .c2r,edx
+       movzx    edx,word[.col2g]
+       shl      edx,ROUND
+       mov      .c2g,edx
+       movzx    edx,word[.col2b]
+       shl      edx,ROUND
+       mov      .c2b,edx
+    .gt_loop2:
+       pushad
+    ; macro .debug
+       mov      edx,.c2r              ; c2r,c2g,c2b,c1r,c1g,c1b - current colors
+       sar      edx,ROUND
+       push     dx
+       mov      edx,.c2g
+       sar      edx,ROUND
+       push     dx
+       mov      edx,.c2b
+       sar      edx,ROUND
+       push     dx
+       sar      ebx,ROUND    ; x2
+       push     bx
+       mov      edx,.c1r
+       sar      edx,ROUND
+       push     dx
+       mov      edx,.c1g
+       sar      edx,ROUND
+       push     dx
+       mov      edx,.c1b
+       sar      edx,ROUND
+       push     dx
+       sar      eax,ROUND
+       push     ax            ; x1
+       push     cx            ; y
+       push     .zz2
+       push     .zz1
+       call     gouraud_line_z
+       popad
+if Ext >= MMX
+       movq     mm0,.c1bM
+       paddd    mm0,qword .dc13bM
+       movq     .c1bM,mm0
+       movq     mm1,.c2bM
+       paddd    mm1,qword .dc23bM
+       movq     .c2bM,mm1
+       movq     mm0,.c1rM
+       paddd    mm0,qword .dc13rM
+       movq     .c1rM,mm0
+       movq     mm1,.c2rM
+       paddd    mm1,qword .dc23rM
+       movq     .c2rM,mm1
+else
+       mov      edx,.dc13r
+       add      .c1r,edx
+       mov      edx,.dc13g
+       add      .c1g,edx
+       mov      edx,.dc13b
+       add      .c1b,edx
+       mov      edx,.dc23r
+       add      .c2r,edx
+       mov      edx,.dc23g
+       add      .c2g,edx
+       mov      edx,.dc23b
+       add      .c2b,edx
+       mov      edx,.dz13
+       add      .zz1,edx
+       mov      edx,.dz23
+       add      .zz2,edx
+end if
+       add      eax,.dx13
+       add      ebx,.dx23
+       inc      cx
+       cmp      cx,.y3
+       jl       .gt_loop2
+   .gt_loop2_end:
+       mov      esp,ebp
+ret 24
+gouraud_line_z:
+;----------------- procedure drawing gouraud line
+;----------------- with z coordinate interpolation
+;----------------- esi - pointer to Z_buffer
+;----------------- edi - pointer to screen buffer
+;----------------- stack:
+.z1  equ dword[ebp+4]   ; z coordiunate shifted left CATMULL_SHIFT
+.z2  equ dword[ebp+8]
+.y   equ word[ebp+12]
+.x1  equ ebp+14
+.c1b equ ebp+16
+.c1g equ ebp+18
+.c1r equ ebp+20
+.x2  equ ebp+22
+.c2b equ ebp+24
+.c2g equ ebp+26
+.c2r equ ebp+28
+.dz   equ dword[ebp-4]
+.dc_b equ dword[ebp-8]
+.dc_g equ dword[ebp-12]
+.dc_r equ dword[ebp-16]
+.c_z  equ dword[ebp-20]
+.cb   equ dword[ebp-24]
+.cg   equ dword[ebp-28]
+.cr   equ dword[ebp-32]
+;.cg2  equ dword[ebp-36]
+.crM  equ ebp-32
+.cgM  equ ebp-28
+.cbM  equ ebp-24
+.dc_rM equ ebp-16
+.dc_gM equ ebp-12
+.dc_bM equ ebp-8
+        mov       ebp,esp
+        mov     ax,.y
+        or      ax,ax
+        jl      .gl_quit
+        mov     bx,[size_y_var]
+        dec     bx
+        cmp     ax,bx ;SIZE_Y
+        jge     .gl_quit
+        mov     eax,dword[.x1]
+        cmp     ax,word[.x2]
+        je      .gl_quit
+        jl      @f
+        xchg    eax,dword[.x2]
+        mov     dword[.x1],eax
+        mov     eax,dword[.c1g]
+        xchg    eax,dword[.c2g]
+        mov     dword[.c1g],eax
+        mov     eax,.z1
+        xchg    eax,.z2
+        mov     .z1,eax
+   @@:
+        mov     bx,[size_x_var]
+        dec     bx
+        cmp     word[.x1],bx  ;SIZE_X
+        jge     .gl_quit
+        cmp     word[.x2],0
+        jle     .gl_quit
+        mov     eax,.z2
+        sub     eax,.z1
+        cdq
+        mov     bx,word[.x2]      ; dz = z2-z1/x2-x1
+        sub     bx,word[.x1]
+        movsx   ebx,bx
+        idiv    ebx
+        push    eax
+        mov     ax,word[.c2b]
+        sub     ax,word[.c1b]
+        cwde
+        shl     eax,ROUND
+        cdq
+        idiv    ebx
+        push    eax
+        mov     ax,word[.c2g]
+        sub     ax,word[.c1g]
+        cwde
+        shl     eax,ROUND
+        cdq
+        idiv    ebx
+        push    eax
+        mov     ax,word[.c2r]
+        sub     ax,word[.c1r]
+        cwde
+        shl     eax,ROUND         ; dc_r = c2r-c1r/x2-x1
+        cdq
+        idiv    ebx
+        push    eax
+        cmp     word[.x1],0     ; clipping on function
+        jg      @f
+        mov     eax,.dz
+        movsx   ebx,word[.x1]
+        neg     ebx
+        imul    ebx
+        add     .z1,eax
+        mov     word[.x1],0
+        mov     eax,.dc_r
+        imul    ebx
+        sar     eax,ROUND
+        add     word[.c1r],ax
+        mov     eax,.dc_g
+        imul    ebx
+        sar     eax,ROUND
+        add     word[.c1g],ax
+        mov     eax,.dc_b
+        imul    ebx
+        sar     eax,ROUND
+        add     word[.c1b],ax
+      @@:
+        mov     bx,[size_x_var]
+        dec     bx
+        cmp     word[.x2],bx  ;SIZE_X
+        jl      @f
+        mov     word[.x2],bx  ;SIZE_X
+     @@:
+        sub     esp,16      ; calculate memory begin
+        movzx   edx,word[size_x_var]  ;SIZE_X       ; in buffers
+        movzx   eax,.y
+        mul     edx
+        movzx   edx,word[.x1]
+        add     eax,edx
+        push    eax
+        lea     eax,[eax*3]
+        add     edi,eax
+        pop     eax
+        shl     eax,2
+        add     esi,eax
+        mov     cx,word[.x2]
+        sub     cx,word[.x1]
+        movzx   ecx,cx
+        mov     ebx,.z1          ; ebx - currrent z shl CATMULL_SIFT
+;if Ext >= SSE
+;        mov     .cz,edx
+;end if
+        mov     edx,.dz          ; edx - delta z
+        movzx   eax,word[.c1r]
+        shl     eax,ROUND
+        mov     .cr,eax
+        movzx   eax,word[.c1g]
+        shl     eax,ROUND
+        mov     .cg,eax
+        movzx   eax,word[.c1b]
+        shl     eax,ROUND
+        mov     .cb,eax
+if Ext = MMX
+;        mov     .c_z,edx
+        movd    mm2,[.dc_bM]         ; delta color blue MMX
+        movd    mm3,[.cbM]           ; current blue MMX
+        movq    mm5,[.dc_rM]
+        movq    mm4,[.crM]
+        pxor    mm6,mm6
+end if
+      .ddraw:
+;if Ext = MMX
+;        movq    mm0,mm3
+;        psrsq   mm0,32
+;        movd    ebx,mm0
+;end if
+        cmp     ebx,dword[esi]   ; esi - z_buffer
+        jge     @f               ; edi - Screen buffer
+if Ext = MMX
+        movq    mm0,mm3          ; mm0, mm1 - temp registers
+        psrld   mm0,ROUND
+        movq    mm1,mm4
+        psrld   mm1,ROUND
+        packssdw  mm1,mm0
+        packuswb  mm1,mm6
+;        movd     [edi],mm1
+        movd      eax,mm1
+        stosw
+        shr       eax,16
+        stosb
+else
+        mov     eax,.cr
+        sar     eax,ROUND
+        stosb
+        mov     eax,.cg
+        sar     eax,ROUND
+        stosb
+        mov     eax,.cb
+        sar     eax,ROUND
+        stosb
+end if
+        mov     dword[esi],ebx
+;if Ext = NON
+        jmp     .no_skip
+;end if
+      @@:
+        add     edi,3
+      .no_skip:
+        add     esi,4
+;if Ext=NON
+        add     ebx,edx
+;end if
+if Ext=MMX
+        paddd   mm3,mm2
+        paddd   mm4,mm5
+else
+        mov     eax,.dc_g
+        add     .cg,eax
+        mov     eax,.dc_b
+        add     .cb,eax
+        mov     eax,.dc_r
+        add     .cr,eax
+end if
+        loop    .ddraw
+   .gl_quit:
+        mov       esp,ebp
+ret 26

 /programs/demos/view3ds/grd_tex.inc
 ,1016 → 1,1016
-CATMULL_SHIFT  equ 8
-ROUND equ 8
-;NON=0
-;MMX=1
-;Ext=MMX
-;TEX_SIZE=0x3fff
-;SIZE_X equ 512
-;SIZE_Y equ 512
-;ROUND = 8
-;TEX_SHIFT equ 6
-; procedure drawing textured triangle with Gouraud shading
-; Z-buffer alghoritm included, Z coord interpolation ----
-; I set the color by this way -- (col1 * col2)/256 ------
-;------------------in - eax - x1 shl 16 + y1 ------------
-;---------------------- ebx - x2 shl 16 + y2 ------------
-;---------------------- ecx - x3 shl 16 + y3 ------------
-;---------------------- esi - pointer to Z-buffer--------
-;---------------------- edx - pointer to texture---------
-;---------------------- Z-buffer filled with dd variables
-;---------------------- shifted CATMULL_SHIFT------------
-;---------------------- edi - pointer to screen buffer---
-;---------------------- stack : colors-------------------
-tex_plus_grd_triangle:
-; parameters :
-   .tex_y3  equ [ebp+38]   ; 36 bytes through stack
-   .tex_x3  equ [ebp+36]
-   .tex_y2  equ [ebp+34]
-   .tex_x2  equ [ebp+32]
-   .tex_y1  equ [ebp+30]
-   .tex_x1  equ [ebp+28]
-   .z3      equ [ebp+26]
-   .col3b   equ [ebp+24]
-   .col3g   equ [ebp+22]
-   .col3r   equ [ebp+20]
-   .z2      equ [ebp+18]
-   .col2b   equ [ebp+16]
-   .col2g   equ [ebp+14]
-   .col2r   equ [ebp+12]
-   .z1      equ [ebp+10]
-   .col1b   equ [ebp+8]
-   .col1g   equ [ebp+6]
-   .col1r   equ [ebp+4]
-; local variables:
-   .tex_ptr   equ dword[ebp-4]
-   .z_ptr     equ dword[ebp-8]
-   .scr_buff  equ dword[ebp-12]
-   .x1        equ  word[ebp-14] ;dw ? ;equ word[ebp-10]
-   .y1        equ  word[ebp-16] ;dw ? ;equ word[ebp-12]
-   .x2        equ  word[ebp-18] ;dw ? ;equ word[ebp-14]
-   .y2        equ  word[ebp-20] ;dw ? ;equ word[ebp-16]
-   .x3        equ  word[ebp-22] ;dw ? ;equ word[ebp-18]
-   .y3        equ  word[ebp-24] ;dw ? ;equ word[ebp-20]
-   .dx12      equ  dword[ebp-28] ;dd ?
-   .tex_dx12  equ  dword[ebp-32] ;dd ?
-   .tex_dy12  equ       [ebp-36] ;dd ?
-   .dz12      equ  dword[ebp-40] ;dd ?
-   .dc12r     equ       [ebp-44] ;dd ?
-   .dc12g     equ  dword[ebp-48] ;dd ?
-   .dc12b     equ       [ebp-52] ;dd ?
-   .dx23      equ  dword[ebp-56] ;dd ?
-   .tex_dx23  equ  dword[ebp-60] ;dd ?
-   .tex_dy23  equ       [ebp-64] ;dd ?
-   .dz23      equ  dword[ebp-68] ;dd ?
-   .dc23r     equ       [ebp-72] ;dd ?
-   .dc23g     equ  dword[ebp-76] ;dd ?
-   .dc23b     equ       [ebp-80] ;dword[ebp-8]dd ?
-   .dx13      equ  dword[ebp-84] ;dd ?
-   .tex_dx13  equ  dword[ebp-88] ;dd ?
-   .tex_dy13  equ       [ebp-92] ;dd ?
-   .dz13      equ  dword[ebp-96] ;dd ?
-   .dc13r     equ       [ebp-100] ;dd ?
-   .dc13g     equ  dword[ebp-104] ;dd ?
-   .dc13b     equ       [ebp-108] ;dd ?
-   .scan_x1   equ  dword[ebp-112] ;dd ?
-   .scan_y1   equ       [ebp-116] ;dd ?
-   .zz1       equ  dword[ebp-120] ;dw ?
-   .cur1r     equ       [ebp-124] ;dw ?
-   .cur1g     equ  dword[ebp-128] ;dw ?
-   .cur1b     equ       [ebp-132] ;dw ?
-   .scan_x2   equ  dword[ebp-136] ;dd ?
-   .scan_y2   equ       [ebp-140] ;dd ?
-   .zz2       equ  dword[ebp-144] ;dw ?
-   .cur2r     equ       [ebp-148] ;dw ?
-   .cur2g     equ  dword[ebp-152] ;dw ?
-   .cur2b     equ       [ebp-156] ;dw ?
-       mov    ebp,esp
-   ;    mov     .tex_ptr,edx
-   ;    mov     .z_ptr,esi
-   ;    mov     .scr_buff,edi
-         push    edx esi edi
-;         push    esi
-;         push    edi
-         mov      edx,80008000h  ; eax,ebx,ecx are ANDd together into edx which means that
-         and      edx,ebx        ; if *all* of them are negative a sign flag is raised
-         and      edx,ecx
-         and      edx,eax
-         test     edx,80008000h  ; Check both X&Y at once
-         jne      .loop2_end
-   .sort3:
-       cmp     ax,bx
-       jle     .sort1
-       xchg    eax,ebx
-if Ext>=MMX
-       movq    mm0, .col1r       ; exchange r, g, b, z
-       movq    mm1, .col2r
-       movq    .col1r ,mm1
-       movq    .col2r ,mm0
-else
-       mov     edx,dword .col1r   ; exchange both r and g
-       xchg    edx,dword .col2r
-       mov     dword .col1r ,edx
-       mov     edx,dword .col1b   ; b and z
-       xchg    edx,dword .col2b
-       mov     dword .col1b ,edx
-end if
-       mov     edx,dword .tex_x1
-       xchg    edx,dword .tex_x2
-       mov     dword .tex_x1 ,edx
- .sort1:
-       cmp      bx,cx
-       jle      .sort2
-       xchg    ebx,ecx
-if Ext>=MMX
-       movq    mm0, .col2r       ; exchange r, g, b, z
-       movq    mm1, .col3r
-       movq    .col3r ,mm0
-       movq    .col2r ,mm1
-else
-       mov     edx,dword .col2r  ; r, g
-       xchg    edx,dword .col3r
-       mov     dword .col2r,edx
-       mov     edx,dword .col2b  ; b, z
-       xchg    edx,dword .col3b
-       mov     dword .col2b,edx
-end if
-       mov     edx,dword .tex_x2
-       xchg    edx,dword .tex_x3
-       mov     dword .tex_x2,edx
-       jmp .sort3
- .sort2:
-       push     eax ebx ecx         ; store in variables
-  ;     push     ebx
-  ;     push     ecx
-;****************** delta computng zone **************
-;+++++++++ first zone
-       mov      bx,.y2       ; calc delta12
-       sub      bx,.y1
-       jnz      .dx12_make
-       mov      ecx,7
-     @@:
-       push     dword 0
-       loop     @b
-       jmp      .dx12_done
-  .dx12_make:
-       mov      ax,.x2
-       sub      ax,.x1
-       cwde
-       movsx    ebx,bx
-       shl      eax,ROUND
-       cdq
-       idiv     ebx
- ;      mov      .dx12,eax
-       push      eax
-if  0  ; Ext=SSE
-       movd      mm0,.col1r    ; 2 words  r, g
-       pxor      mm1,mm1
-       punpcklwd mm0,mm1
-       cvtpi2ps  xmm0,mm0
-       movlhps   xmm0,xmm0
-       movd      mm0,.col1g    ; 2 words  b, z
-       punpcklwd mm0,mm1
-       cvtpi2ps  xmm0,mm0
-       ; xmm0=four float double words
-       divss     xmm0,.pack3
-       ;convert and insert mm0 to lower xmm1     ..
-end if
-       mov      ax,word .tex_x2
-       sub      ax,word .tex_x1
-       cwde
-       shl      eax,ROUND
-       cdq
-       idiv     ebx
-;       mov      .tex_dx12r,eax
-       push       eax
-       mov        ax,word .tex_y2
-       sub        ax,word .tex_y1
-       cwde
-       shl      eax,ROUND
-       cdq
-       idiv     ebx
-;       mov      .tex_dx12,eax
-       push     eax
-       mov      ax,word .z2
-       sub      ax,word .z1
-       cwde
-       shl      eax,CATMULL_SHIFT
-       cdq
-       idiv     ebx
- ;      mov      .dz12,eax
-       push     eax       ; .dza12
-       mov      ax,word .col2r
-       sub      ax,word .col1r
-       cwde
-       shl      eax,ROUND
-       cdq
-       idiv     ebx
- ;      mov      .dc12r,eax
-       push       eax
-       mov        ax,word .col2g
-       sub        ax,word .col1g
-       cwde
-       shl      eax,ROUND
-       cdq
-       idiv     ebx
- ;      mov      .dc12g,eax
-       push     eax
-       mov      ax,word .col2b         ;;---
-       sub      ax,word .col1b
-       cwde
-       shl      eax,ROUND
-       cdq
-       idiv     ebx
-;       mov      .dc12b,eax
-       push     eax
-;+++++++++++++++++ second zone +++++++++++++
-   .dx12_done:
-       mov      bx,.y3       ; calc delta23
-       sub      bx,.y2
-       jnz      .dx23_make
-       mov      ecx,7
-    @@:
-       push     dword 0
-       loop     @b
-       jmp      .dx23_done
-  .dx23_make:
-       mov      ax,.x3
-       sub      ax,.x2
-       cwde
-       movsx    ebx,bx
-       shl      eax,ROUND
-       cdq
-       idiv     ebx
- ;      mov      .dx23,eax
-       push      eax
-       mov      ax,word .tex_x3
-       sub      ax,word .tex_x2
-       cwde
-       shl      eax,ROUND
-       cdq
-       idiv     ebx
-;       mov      .tex_dx23,eax
-       push       eax
-       mov        ax,word .tex_y3
-       sub        ax,word .tex_y2
-       cwde
-       shl      eax,ROUND
-       cdq
-       idiv     ebx
-;       mov      .tex_dy23,eax
-       push     eax
-       mov      ax,word .z3
-       sub      ax,word .z2
-       cwde                                ;
-       shl      eax,CATMULL_SHIFT          ;  2222222
-       cdq                                 ;  2     2
-       idiv     ebx                        ;        2
-;       mov      .dz23,eax                 ;       2
-       push     eax       ; .dza12         ;      2
-                                           ;     2
-       mov      ax,word .col3r             ;    2
-       sub      ax,word .col2r             ;   2222222
-       cwde                                ;  second delta
-       shl      eax,ROUND                  ;
-       cdq                                 ;
-       idiv     ebx                        ;
-;       mov      .dc23r,eax                 ;
-       push       eax
-       mov        ax,word .col3g
-       sub        ax,word .col2g
-       cwde
-       shl      eax,ROUND
-       cdq
-       idiv     ebx
-;       mov      .dc23g,eax
-       push     eax
-       mov      ax,word .col3b         ;;---
-       sub      ax,word .col2b
-       cwde
-       shl      eax,ROUND
-       cdq
-       idiv     ebx
-;       mov      .dc23b,eax
-       push     eax
-   .dx23_done:
-;++++++++++++++++++third zone++++++++++++++++++++++++
-       mov      bx,.y3       ; calc delta13
-       sub      bx,.y1
-       jnz      .dx13_make
-       mov      ecx,7
-     @@:
-       push     dword 0
-       loop     @b
-       jmp      .dx13_done
-  .dx13_make:
-       mov      ax,.x3
-       sub      ax,.x1
-       cwde
-       movsx    ebx,bx
-       shl      eax,ROUND
-       cdq
-       idiv     ebx
-;       mov      .dx13,eax
-       push      eax
-       mov      ax,word .tex_x3            ; triangle b
-       sub      ax,word .tex_x1
-       cwde
-       shl      eax,ROUND
-       cdq
-       idiv     ebx
-;       mov      .tex_dx13r,eax
-       push       eax
-       mov        ax,word .tex_y3
-       sub        ax,word .tex_y1
-       cwde
-       shl      eax,ROUND
-       cdq
-       idiv     ebx
-;       mov      .tex_dy13,eax
-       push     eax
-       mov      ax,word .z3
-       sub      ax,word .z1            ;    333333333
-       cwde                            ;   3        3
-       shl      eax,CATMULL_SHIFT      ;            3
-       cdq                             ;            3
-       idiv     ebx                    ;            3
-;       mov      .dz13,eax             ;            3
-       push     eax       ; .dza12     ;            3
-                                       ;            3
-       mov      ax,word .col3r         ;    3333333333
-       sub      ax,word .col1r         ;            3
-       cwde                            ;            3
-       shl      eax,ROUND              ;            3
-       cdq                             ;            3
-       idiv     ebx                    ;            3
-  ;    mov      .dc13r,eax             ;    3       3
-       push       eax                 ;     33333333
-       mov        ax,word .col3g
-       sub        ax,word .col1g
-       cwde
-       shl      eax,ROUND
-       cdq
-       idiv     ebx
- ;      mov      .dc13g,eax
-       push     eax
-       mov      ax,word .col3b         ;;---
-       sub      ax,word .col1b
-       cwde
-       shl      eax,ROUND
-       cdq
-       idiv     ebx
-;       mov      .dc13b,eax
-       push     eax
-   .dx13_done:
-; <<<<<<<  ::delta zone end+++++++++++++++++++++ >>>>>>>>
-       sub      esp,55   ;(12*4)
-       movsx    eax,.x1                    ; eax - cur x1
-       shl      eax,ROUND                  ; ebx - cur x2
-       mov      ebx,eax
-       movsx    edx,word .z1
-       shl      edx,CATMULL_SHIFT
-       mov      .zz1,edx
-       mov      .zz2,edx
-       movzx    edi,word .col1r
-       shl      edi,ROUND
-       mov      .cur1r,edi
-       mov      .cur2r,edi
-       movzx    esi,word .col1g
-       shl      esi,ROUND
-       mov      .cur1g,esi
-       mov      .cur2g,esi
-       movzx    edx,word .col1b
-       shl      edx,ROUND
-       mov      .cur1b,edx
-       mov      .cur2b,edx
-       movzx    edi,word .tex_x1
-       shl      edi,ROUND
-       mov      .scan_x1,edi
-       mov      .scan_x2,edi
-       movzx    edx,word .tex_y1
-       shl      edx,ROUND
-       mov      .scan_y1,edx
-       mov      .scan_y2,edx
-       mov      cx,.y1
-       cmp      cx,.y2
-       jge      .loop1_end
-    .loop_1:
-    ;   push     eax ebx ebp
-       pushad
-       push     .tex_ptr
-       push     .scr_buff
-       push     .z_ptr
-       push     cx
-       push     .zz2
-       push     .scan_x2
-       push     dword .scan_y2
-       push     dword .cur2r
-       push     .cur2g
-       push     dword .cur2b
-       push     .zz1
-       push     .scan_x1
-       push     dword .scan_y1
-       push     dword .cur1r
-       push     .cur1g
-       push     dword .cur1b
-       sar      eax,ROUND
-       sar      ebx,ROUND
-       call     horizontal_tex_grd_line
-      ; pop      ebp ebx eax
-       popad
-if (Ext = MMX)|(Ext=SSE)
-       movq     mm0,.cur1b
-       movq     mm1,.cur1r
-       movq     mm2,.scan_y1
-       movq     mm3,.cur2b
-       movq     mm4,.cur2r
-       movq     mm5,.scan_y2
-       paddd    mm0,.dc13b
-       paddd    mm1,.dc13r
-       paddd    mm2,.tex_dy13
-       paddd    mm3,.dc12b
-       paddd    mm4,.dc12r
-       paddd    mm5,.tex_dy12
-       movq     .cur1b,mm0
-       movq     .cur1r,mm1
-       movq     .scan_y1,mm2
-       movq     .cur2b,mm3
-       movq     .cur2r,mm4
-       movq     .scan_y2,mm5
-end if
-if Ext >= SSE2
-       movups   xmm0,.cur1b
-       movups   xmm1,.dc13b
-       movups   xmm2,.cur2b
-       movups   xmm3,.dc12b
-       movq     mm2,.scan_y1
-       movq     mm5,.scan_y2
-       paddd    xmm0,xmm1
-       paddd    xmm2,xmm3
-       paddd    mm2,.tex_dy13
-       paddd    mm5,.tex_dy12
-       movq     .scan_y1,mm2
-       movq     .scan_y2,mm5
-       movups   .cur1b,xmm0
-       movups   .cur2b,xmm2
-end if
-if Ext = NON
-       mov      edx,.dc13b
-       add      .cur1b,edx
-       mov      esi,.dc13g
-       add      .cur1g,esi
-       mov      edi,.dc13r
-       add      .cur1r,edi
-       mov      edx,.dz13
-       add      .zz1,edx
-       mov      edx,.tex_dx13
-       add      .scan_x1,edx
-       mov      esi,.tex_dy13
-       add      .scan_y1,esi
-       mov      edi,.dc12b
-       add      .cur2b,edi
-       mov      esi,.dc12g
-       add      .cur2g,esi
-       mov      edx,.dc12r
-       add      .cur2r,edx
-       mov      edi,.tex_dx12
-       add      .scan_x2,edi
-       mov      esi,.tex_dy12
-       add      .scan_y2,esi
-       mov      edx,.dz12
-       add      .zz2,edx
-end if
-       add      eax,.dx13
-       add      ebx,.dx12
-       inc      cx
-       cmp      cx,.y2
-       jl      .loop_1
- .loop1_end:
-       movzx    ecx,.y2
-       cmp      cx,.y3
-       jge      .loop2_end
-       movsx    ebx,.x2                    ; eax - cur x1
-       shl      ebx,ROUND                  ; ebx - cur x2
-       movsx    edx,word .z2
-       shl      edx,CATMULL_SHIFT
-;       mov      .zz1,edx
-       mov      .zz2,edx
-       movzx    edi,word .col2r
-       shl      edi,ROUND
-   ;    mov      .cur1r,edi
-       mov      .cur2r,edi
-       movzx    esi,word .col2g
-       shl      esi,ROUND
-   ;    mov      .cur1g,esi
-       mov      .cur2g,esi
-       movzx    edx,word .col2b
-       shl      edx,ROUND
-   ;    mov      .cur1b,edx
-       mov      .cur2b,edx
-       movzx    edi,word .tex_x2
-       shl      edi,ROUND
- ;      mov      .scan_x1,edi
-       mov      .scan_x2,edi
-       movzx    edx,word .tex_y2
-       shl      edx,ROUND
- ;      mov      .scan_y1,edx
-       mov      .scan_y2,edx
-  .loop_2:
-       pushad
-       push     .tex_ptr
-       push     .scr_buff
-       push     .z_ptr
-       push     cx
-       push     .zz2
-       push     .scan_x2
-       push     dword .scan_y2
-       push     dword .cur2r
-       push     .cur2g
-       push     dword .cur2b
-       push     .zz1
-       push     .scan_x1
-       push     dword .scan_y1
-       push     dword .cur1r
-       push     .cur1g
-       push     dword .cur1b
-       sar      eax,ROUND
-       sar      ebx,ROUND
-       call     horizontal_tex_grd_line
-       popad
-if (Ext = MMX)|(Ext=SSE)
-       movq     mm0,.cur1b
-       movq     mm1,.cur1r
-       movq     mm2,.scan_y1
-       movq     mm3,.cur2b
-       movq     mm4,.cur2r
-       movq     mm5,.scan_y2
-       paddd    mm0,.dc13b
-       paddd    mm1,.dc13r
-       paddd    mm2,.tex_dy13
-       paddd    mm3,.dc23b
-       paddd    mm4,.dc23r
-       paddd    mm5,.tex_dy23
-       movq     .cur1b,mm0
-       movq     .cur1r,mm1
-       movq     .scan_y1,mm2
-       movq     .cur2b,mm3
-       movq     .cur2r,mm4
-       movq     .scan_y2,mm5
-end if
-if Ext >= SSE2
-       movups   xmm0,.cur1b
-       movups   xmm1,.dc13b
-       movups   xmm2,.cur2b
-       movups   xmm3,.dc23b
-       movq     mm2,.scan_y1
-       movq     mm5,.scan_y2
-       paddd    xmm0,xmm1
-       paddd    xmm2,xmm3
-       paddd    mm2,.tex_dy13
-       paddd    mm5,.tex_dy23
-       movq     .scan_y1,mm2
-       movq     .scan_y2,mm5
-       movups   .cur1b,xmm0
-       movups   .cur2b,xmm2
-end if
-if Ext = NON
-       mov      edx,.dc13b
-       add      .cur1b,edx
-       mov      esi,.dc13g
-       add      .cur1g,esi
-       mov      edi,.dc13r
-       add      .cur1r,edi
-       mov      edx,.tex_dx13
-       add      .scan_x1,edx
-       mov      esi,.tex_dy13
-       add      .scan_y1,esi
-       mov      edx,.dz13
-       add      .zz1,edx
-       mov      edi,.dc23b
-       add      .cur2b,edi
-       mov      esi,.dc23g
-       add      .cur2g,esi
-       mov      edx,.dc23r
-       add      .cur2r,edx
-       mov      edi,.tex_dx23
-       add      .scan_x2,edi
-       mov      esi,.tex_dy23
-       add      .scan_y2,esi
-       mov      edx,.dz23
-       add      .zz2,edx
-end if
-       add      eax,.dx13
-       add      ebx,.dx23
-       inc      cx
-       cmp      cx,.y3
-       jl       .loop_2
-.loop2_end:
-       mov      esp,ebp
-ret 36
-horizontal_tex_grd_line:
-;in:
-; eax : x1, ebx : x2
-.tex_ptr  equ [ebp+62]
-.screen   equ [ebp+58]
-.z_buffer equ [ebp+54]
-.y        equ [ebp+52]
-.z2     equ [ebp+48]
-.tex_x2 equ [ebp+44]
-.tex_y2 equ [ebp+40]
-.r2     equ [ebp+36]
-.g2     equ [ebp+32]
-.b2     equ [ebp+28]
-.z1     equ [ebp+24]
-.tex_x1 equ [ebp+20]
-.tex_y1 equ [ebp+16]
-.r1     equ [ebp+12]
-.g1     equ [ebp+8]
-.b1     equ [ebp+4]
-.x1 equ  word[ebp-2]
-.x2 equ  word[ebp-4]
-.dz equ dword[ebp-8]
-.db equ dword[ebp-12]
-.dg equ dword[ebp-16]
-.dr equ dword[ebp-20]
-.dtex_x equ dword[ebp-24]
-.dtex_y equ dword[ebp-28]
-.c_ty equ [ebp-32]
-.c_tx equ [ebp-36]
-.cb  equ  [ebp-40]
-.cg  equ  [ebp-44]
-.cr  equ  [ebp-48]
-.t_col equ [ebp-52]
-.dtex_yM equ qword[ebp-28]
-.drM equ qword[ebp-20]
-.dbM equ qword[ebp-12]
-        mov     ebp,esp
-  ;      sub     esp,30
-        mov     cx,word .y
-        or      cx,cx
-        jl      .quit_l
-        cmp     cx,word[size_y_var]  ;SIZE_Y
-        jge     .quit_l
-        cmp     ax,bx
-        je      .quit_l
-        jl      @f
-        xchg    eax,ebx
-if Ext=NON
-        mov     ecx,dword .r1
-        xchg    ecx, .r2
-        mov     dword .r1, ecx
-        mov     ecx,dword .g1
-        xchg    ecx, .g2
-        mov     dword .g1, ecx
-        mov     ecx,dword .b1
-        xchg    ecx, .b2
-        mov     dword .b1, ecx
-        mov     ecx,dword .tex_x1
-        xchg    ecx, .tex_x2
-        mov     dword .tex_x1, ecx
-        mov     ecx,dword .tex_y1
-        xchg    ecx, .tex_y2
-        mov     dword .tex_y1, ecx
-        mov     ecx,dword .z1
-        xchg    ecx, .z2
-        mov     dword .z1, ecx
-end if
-if (Ext=MMX)
-        movq    mm0,.b1        ; b, g
-        movq    mm1,.b2
-        movq    .b1, mm1
-        movq    .b2, mm0
-        movq    mm2,.r1        ; r, y
-        movq    mm3,.r2
-        movq    .r1,mm3
-        movq    .r2,mm2
-        movq    mm4,.tex_x1    ; x, z
-        movq    mm5,.tex_x2
-        movq    .tex_x1,mm5
-        movq    .tex_x2,mm4
-end if
-if Ext>=SSE
-       movups  xmm0,.b1
-       movups  xmm1,.b2
-       movups  .b1,xmm1
-       movups  .b2,xmm0
-       movq    mm4,.tex_x1    ; x, z
-       movq    mm5,.tex_x2
-       movq    .tex_x1,mm5
-       movq    .tex_x2,mm4
-end if
-    @@:
-        or      bx,bx
-        jle     .quit_l
-        cmp     ax,word[size_x_var]  ;SIZE_X
-        jge     .quit_l
-        push    ax
-        push    bx
-        mov     eax,.z2            ; delta zone************
-        sub     eax,.z1
-        cdq
-        mov     bx,.x2
-        sub     bx,.x1
-        movsx   ebx,bx
-        idiv    ebx
-        push    eax  ; .dz
-        mov     eax,.b2
-        sub     eax,.b1
-        cdq
-        idiv    ebx
-        push    eax  ; .db
-        mov     eax,.g2
-        sub     eax,.g1
-        cdq
-        idiv    ebx
-        push    eax  ; .dg
-        mov     eax,.r2
-        sub     eax,.r1
-        cdq
-        idiv    ebx
-        push    eax  ; .dr
-        mov     eax,.tex_x2
-        sub     eax,.tex_x1
-        cdq
-        idiv    ebx
-        push    eax  ; .dtex_x
-        mov     eax,.tex_y2
-        sub     eax,.tex_y1
-        cdq
-        idiv    ebx
-        push    eax  ; .dtey_x
-        cmp     .x1,0
-        jg      @f
-        mov     eax,.dz     ; clipping
-        movsx   ebx,.x1
-        neg     ebx
-        imul    ebx
-        add     .z1,eax
-        mov     .x1,0
-        mov    eax,.dr
-        imul   ebx
-        add    .r1,eax
-;if  Ext=NON
-        mov    eax,.dg
-        imul   ebx
-        add    .g1,eax
-        mov    eax,.db
-        imul   ebx
-        add    .b1,eax
-        mov    eax,.dtex_x
-        imul   ebx
-        add    .tex_x1,eax
-        mov    eax,.dtex_y
-        imul   ebx
-        add    .tex_y1,eax
-   @@:
-        movsx     edx,word[size_x_var]  ;SIZE_X
-        cmp       .x2,dx
-        jl        @f
-        mov       .x2,dx
-    @@:
-; calc line addres begin in screen and Z buffer
-        movsx     eax,word .y
-        mul       edx
-        movsx     edx,.x1
-        add       eax,edx
-        mov       esi,eax
-        shl       esi,2
-        add       esi,.z_buffer
-        lea       eax,[eax*3]
-        mov       edi,.screen
-        add       edi,eax
-        mov       cx,.x2
-        sub       cx,.x1
-        movzx     ecx,cx
-; init current variables
-        push      dword .tex_y1
-;if  Ext=NON
-        push      dword .tex_x1
-        push      dword .b1
-        push      dword .g1
-        push      dword .r1
-if Ext>=MMX
-        movq      mm4,.cr  ; lo -> r,g
-        movq      mm6,.cb  ; hi -> b, tex_x
-        pxor      mm0,mm0
-end if
-        mov       ebx,.z1
-      .ddraw:
-        cmp       ebx,dword[esi]
-        jge       @f
-        mov       eax,.c_ty
-;  if ROUND<TEX_SHIFT
-;        shl        eax,TEX_SHIFT-ROUND
-;  end if
-;  if ROUND>TEX_SHIFT
-;        shr        eax,ROUND-TEX_SHIFT
-;  end if
-        shr       eax,ROUND
-        shl       Eax,TEX_SHIFT
-        mov       edx,.c_tx   ; calc texture pixel mem addres
-        shr       edx,ROUND
-        add       eax,edx
-        and       eax,TEXTURE_SIZE   ; cutting
-        lea       eax,[3*eax]
-        add       eax,.tex_ptr
-        mov       dword[esi],ebx
-if    Ext = NON
-        mov       eax,dword[eax]
- ;      mov        .tex_col,eax
-        push      ax
-        shl       eax,8
-        pop       ax
-        mov       edx,.cr
-        sar       edx,ROUND
-        mul       dl         ; al*dl
-        shr       ax,8
-        stosb
-        ror       eax,16
-        push      ax
-        mov       edx,.cg
-        sar       edx,ROUND
-        mul       dl
-        shr       ax,8
-        stosb
-        pop       ax
-        shr       ax,8
-        mov       edx,.cb
-        sar       edx,ROUND
-        mul       dl
-        shr       ax,8
-        stosb
-        jmp       .no_skip
-else
-        movd      mm1,[eax]
-        punpcklbw mm1,mm0
-        movq      mm3,mm4  ;.cr     ; lo -> r,g
-        movq      mm5,mm6  ;.cb     ; lo -> b,tex_x
-        psrld     mm3,ROUND  ;
-        psrld     mm5,ROUND  ;
-        packssdw  mm3,mm5
-        pmullw    mm1,mm3
-        psrlw     mm1,8
-        packuswb  mm1,mm0
-        movd      [edi],mm1
-end if
-       mov        dword[esi],ebx
-if Ext = NON
-       jmp        .no_skip
-end if
-     @@:
-       add        edi,3
-     .no_skip:
-       add        esi,4
-       add        ebx,.dz
-       mov        eax,.dtex_x
-       add        .c_tx, eax
-       mov        edx,.dtex_y
-       add        .c_ty, edx
-if Ext=NON
-       mov        eax,.dr
-       add        .cr,eax
-       mov        edx,.dg
-       add        .cg,edx
-       mov        eax,.db
-       add        .cb,eax
-else
-        paddd      mm4,.drM
-        paddd      mm6,.dbM
-;;        paddd      mm7,.dtex_y      ; mm4 - b, g
-;;        movq       .c_tx,mm7
-                                    ; mm6 - r, x
-end if                              ; mm7 - y, x
-       dec      ecx
-       jnz      .ddraw
-  .quit_l:
-        mov     esp,ebp
-ret 42+20  ; horizontal line
+CATMULL_SHIFT  equ 8
+ROUND equ 8
+;NON=0
+;MMX=1
+;Ext=MMX
+;TEX_SIZE=0x3fff
+;SIZE_X equ 512
+;SIZE_Y equ 512
+;ROUND = 8
+;TEX_SHIFT equ 6
+; procedure drawing textured triangle with Gouraud shading
+; Z-buffer alghoritm included, Z coord interpolation ----
+; I set the color by this way -- (col1 * col2)/256 ------
+;------------------in - eax - x1 shl 16 + y1 ------------
+;---------------------- ebx - x2 shl 16 + y2 ------------
+;---------------------- ecx - x3 shl 16 + y3 ------------
+;---------------------- esi - pointer to Z-buffer--------
+;---------------------- edx - pointer to texture---------
+;---------------------- Z-buffer filled with dd variables
+;---------------------- shifted CATMULL_SHIFT------------
+;---------------------- edi - pointer to screen buffer---
+;---------------------- stack : colors-------------------
+tex_plus_grd_triangle:
+; parameters :
+   .tex_y3  equ [ebp+38]   ; 36 bytes through stack
+   .tex_x3  equ [ebp+36]
+   .tex_y2  equ [ebp+34]
+   .tex_x2  equ [ebp+32]
+   .tex_y1  equ [ebp+30]
+   .tex_x1  equ [ebp+28]
+   .z3      equ [ebp+26]
+   .col3b   equ [ebp+24]
+   .col3g   equ [ebp+22]
+   .col3r   equ [ebp+20]
+   .z2      equ [ebp+18]
+   .col2b   equ [ebp+16]
+   .col2g   equ [ebp+14]
+   .col2r   equ [ebp+12]
+   .z1      equ [ebp+10]
+   .col1b   equ [ebp+8]
+   .col1g   equ [ebp+6]
+   .col1r   equ [ebp+4]
+; local variables:
+   .tex_ptr   equ dword[ebp-4]
+   .z_ptr     equ dword[ebp-8]
+   .scr_buff  equ dword[ebp-12]
+   .x1        equ  word[ebp-14] ;dw ? ;equ word[ebp-10]
+   .y1        equ  word[ebp-16] ;dw ? ;equ word[ebp-12]
+   .x2        equ  word[ebp-18] ;dw ? ;equ word[ebp-14]
+   .y2        equ  word[ebp-20] ;dw ? ;equ word[ebp-16]
+   .x3        equ  word[ebp-22] ;dw ? ;equ word[ebp-18]
+   .y3        equ  word[ebp-24] ;dw ? ;equ word[ebp-20]
+   .dx12      equ  dword[ebp-28] ;dd ?
+   .tex_dx12  equ  dword[ebp-32] ;dd ?
+   .tex_dy12  equ       [ebp-36] ;dd ?
+   .dz12      equ  dword[ebp-40] ;dd ?
+   .dc12r     equ       [ebp-44] ;dd ?
+   .dc12g     equ  dword[ebp-48] ;dd ?
+   .dc12b     equ       [ebp-52] ;dd ?
+   .dx23      equ  dword[ebp-56] ;dd ?
+   .tex_dx23  equ  dword[ebp-60] ;dd ?
+   .tex_dy23  equ       [ebp-64] ;dd ?
+   .dz23      equ  dword[ebp-68] ;dd ?
+   .dc23r     equ       [ebp-72] ;dd ?
+   .dc23g     equ  dword[ebp-76] ;dd ?
+   .dc23b     equ       [ebp-80] ;dword[ebp-8]dd ?
+   .dx13      equ  dword[ebp-84] ;dd ?
+   .tex_dx13  equ  dword[ebp-88] ;dd ?
+   .tex_dy13  equ       [ebp-92] ;dd ?
+   .dz13      equ  dword[ebp-96] ;dd ?
+   .dc13r     equ       [ebp-100] ;dd ?
+   .dc13g     equ  dword[ebp-104] ;dd ?
+   .dc13b     equ       [ebp-108] ;dd ?
+   .scan_x1   equ  dword[ebp-112] ;dd ?
+   .scan_y1   equ       [ebp-116] ;dd ?
+   .zz1       equ  dword[ebp-120] ;dw ?
+   .cur1r     equ       [ebp-124] ;dw ?
+   .cur1g     equ  dword[ebp-128] ;dw ?
+   .cur1b     equ       [ebp-132] ;dw ?
+   .scan_x2   equ  dword[ebp-136] ;dd ?
+   .scan_y2   equ       [ebp-140] ;dd ?
+   .zz2       equ  dword[ebp-144] ;dw ?
+   .cur2r     equ       [ebp-148] ;dw ?
+   .cur2g     equ  dword[ebp-152] ;dw ?
+   .cur2b     equ       [ebp-156] ;dw ?
+       mov    ebp,esp
+   ;    mov     .tex_ptr,edx
+   ;    mov     .z_ptr,esi
+   ;    mov     .scr_buff,edi
+         push    edx esi edi
+;         push    esi
+;         push    edi
+         mov      edx,80008000h  ; eax,ebx,ecx are ANDd together into edx which means that
+         and      edx,ebx        ; if *all* of them are negative a sign flag is raised
+         and      edx,ecx
+         and      edx,eax
+         test     edx,80008000h  ; Check both X&Y at once
+         jne      .loop2_end
+   .sort3:
+       cmp     ax,bx
+       jle     .sort1
+       xchg    eax,ebx
+if Ext>=MMX
+       movq    mm0, .col1r       ; exchange r, g, b, z
+       movq    mm1, .col2r
+       movq    .col1r ,mm1
+       movq    .col2r ,mm0
+else
+       mov     edx,dword .col1r   ; exchange both r and g
+       xchg    edx,dword .col2r
+       mov     dword .col1r ,edx
+       mov     edx,dword .col1b   ; b and z
+       xchg    edx,dword .col2b
+       mov     dword .col1b ,edx
+end if
+       mov     edx,dword .tex_x1
+       xchg    edx,dword .tex_x2
+       mov     dword .tex_x1 ,edx
+ .sort1:
+       cmp      bx,cx
+       jle      .sort2
+       xchg    ebx,ecx
+if Ext>=MMX
+       movq    mm0, .col2r       ; exchange r, g, b, z
+       movq    mm1, .col3r
+       movq    .col3r ,mm0
+       movq    .col2r ,mm1
+else
+       mov     edx,dword .col2r  ; r, g
+       xchg    edx,dword .col3r
+       mov     dword .col2r,edx
+       mov     edx,dword .col2b  ; b, z
+       xchg    edx,dword .col3b
+       mov     dword .col2b,edx
+end if
+       mov     edx,dword .tex_x2
+       xchg    edx,dword .tex_x3
+       mov     dword .tex_x2,edx
+       jmp .sort3
+ .sort2:
+       push     eax ebx ecx         ; store in variables
+  ;     push     ebx
+  ;     push     ecx
+;****************** delta computng zone **************
+;+++++++++ first zone
+       mov      bx,.y2       ; calc delta12
+       sub      bx,.y1
+       jnz      .dx12_make
+       mov      ecx,7
+     @@:
+       push     dword 0
+       loop     @b
+       jmp      .dx12_done
+  .dx12_make:
+       mov      ax,.x2
+       sub      ax,.x1
+       cwde
+       movsx    ebx,bx
+       shl      eax,ROUND
+       cdq
+       idiv     ebx
+ ;      mov      .dx12,eax
+       push      eax
+if  0  ; Ext=SSE
+       movd      mm0,.col1r    ; 2 words  r, g
+       pxor      mm1,mm1
+       punpcklwd mm0,mm1
+       cvtpi2ps  xmm0,mm0
+       movlhps   xmm0,xmm0
+       movd      mm0,.col1g    ; 2 words  b, z
+       punpcklwd mm0,mm1
+       cvtpi2ps  xmm0,mm0
+       ; xmm0=four float double words
+       divss     xmm0,.pack3
+       ;convert and insert mm0 to lower xmm1     ..
+end if
+       mov      ax,word .tex_x2
+       sub      ax,word .tex_x1
+       cwde
+       shl      eax,ROUND
+       cdq
+       idiv     ebx
+;       mov      .tex_dx12r,eax
+       push       eax
+       mov        ax,word .tex_y2
+       sub        ax,word .tex_y1
+       cwde
+       shl      eax,ROUND
+       cdq
+       idiv     ebx
+;       mov      .tex_dx12,eax
+       push     eax
+       mov      ax,word .z2
+       sub      ax,word .z1
+       cwde
+       shl      eax,CATMULL_SHIFT
+       cdq
+       idiv     ebx
+ ;      mov      .dz12,eax
+       push     eax       ; .dza12
+       mov      ax,word .col2r
+       sub      ax,word .col1r
+       cwde
+       shl      eax,ROUND
+       cdq
+       idiv     ebx
+ ;      mov      .dc12r,eax
+       push       eax
+       mov        ax,word .col2g
+       sub        ax,word .col1g
+       cwde
+       shl      eax,ROUND
+       cdq
+       idiv     ebx
+ ;      mov      .dc12g,eax
+       push     eax
+       mov      ax,word .col2b         ;;---
+       sub      ax,word .col1b
+       cwde
+       shl      eax,ROUND
+       cdq
+       idiv     ebx
+;       mov      .dc12b,eax
+       push     eax
+;+++++++++++++++++ second zone +++++++++++++
+   .dx12_done:
+       mov      bx,.y3       ; calc delta23
+       sub      bx,.y2
+       jnz      .dx23_make
+       mov      ecx,7
+    @@:
+       push     dword 0
+       loop     @b
+       jmp      .dx23_done
+  .dx23_make:
+       mov      ax,.x3
+       sub      ax,.x2
+       cwde
+       movsx    ebx,bx
+       shl      eax,ROUND
+       cdq
+       idiv     ebx
+ ;      mov      .dx23,eax
+       push      eax
+       mov      ax,word .tex_x3
+       sub      ax,word .tex_x2
+       cwde
+       shl      eax,ROUND
+       cdq
+       idiv     ebx
+;       mov      .tex_dx23,eax
+       push       eax
+       mov        ax,word .tex_y3
+       sub        ax,word .tex_y2
+       cwde
+       shl      eax,ROUND
+       cdq
+       idiv     ebx
+;       mov      .tex_dy23,eax
+       push     eax
+       mov      ax,word .z3
+       sub      ax,word .z2
+       cwde                                ;
+       shl      eax,CATMULL_SHIFT          ;  2222222
+       cdq                                 ;  2     2
+       idiv     ebx                        ;        2
+;       mov      .dz23,eax                 ;       2
+       push     eax       ; .dza12         ;      2
+                                           ;     2
+       mov      ax,word .col3r             ;    2
+       sub      ax,word .col2r             ;   2222222
+       cwde                                ;  second delta
+       shl      eax,ROUND                  ;
+       cdq                                 ;
+       idiv     ebx                        ;
+;       mov      .dc23r,eax                 ;
+       push       eax
+       mov        ax,word .col3g
+       sub        ax,word .col2g
+       cwde
+       shl      eax,ROUND
+       cdq
+       idiv     ebx
+;       mov      .dc23g,eax
+       push     eax
+       mov      ax,word .col3b         ;;---
+       sub      ax,word .col2b
+       cwde
+       shl      eax,ROUND
+       cdq
+       idiv     ebx
+;       mov      .dc23b,eax
+       push     eax
+   .dx23_done:
+;++++++++++++++++++third zone++++++++++++++++++++++++
+       mov      bx,.y3       ; calc delta13
+       sub      bx,.y1
+       jnz      .dx13_make
+       mov      ecx,7
+     @@:
+       push     dword 0
+       loop     @b
+       jmp      .dx13_done
+  .dx13_make:
+       mov      ax,.x3
+       sub      ax,.x1
+       cwde
+       movsx    ebx,bx
+       shl      eax,ROUND
+       cdq
+       idiv     ebx
+;       mov      .dx13,eax
+       push      eax
+       mov      ax,word .tex_x3            ; triangle b
+       sub      ax,word .tex_x1
+       cwde
+       shl      eax,ROUND
+       cdq
+       idiv     ebx
+;       mov      .tex_dx13r,eax
+       push       eax
+       mov        ax,word .tex_y3
+       sub        ax,word .tex_y1
+       cwde
+       shl      eax,ROUND
+       cdq
+       idiv     ebx
+;       mov      .tex_dy13,eax
+       push     eax
+       mov      ax,word .z3
+       sub      ax,word .z1            ;    333333333
+       cwde                            ;   3        3
+       shl      eax,CATMULL_SHIFT      ;            3
+       cdq                             ;            3
+       idiv     ebx                    ;            3
+;       mov      .dz13,eax             ;            3
+       push     eax       ; .dza12     ;            3
+                                       ;            3
+       mov      ax,word .col3r         ;    3333333333
+       sub      ax,word .col1r         ;            3
+       cwde                            ;            3
+       shl      eax,ROUND              ;            3
+       cdq                             ;            3
+       idiv     ebx                    ;            3
+  ;    mov      .dc13r,eax             ;    3       3
+       push       eax                 ;     33333333
+       mov        ax,word .col3g
+       sub        ax,word .col1g
+       cwde
+       shl      eax,ROUND
+       cdq
+       idiv     ebx
+ ;      mov      .dc13g,eax
+       push     eax
+       mov      ax,word .col3b         ;;---
+       sub      ax,word .col1b
+       cwde
+       shl      eax,ROUND
+       cdq
+       idiv     ebx
+;       mov      .dc13b,eax
+       push     eax
+   .dx13_done:
+; <<<<<<<  ::delta zone end+++++++++++++++++++++ >>>>>>>>
+       sub      esp,55   ;(12*4)
+       movsx    eax,.x1                    ; eax - cur x1
+       shl      eax,ROUND                  ; ebx - cur x2
+       mov      ebx,eax
+       movsx    edx,word .z1
+       shl      edx,CATMULL_SHIFT
+       mov      .zz1,edx
+       mov      .zz2,edx
+       movzx    edi,word .col1r
+       shl      edi,ROUND
+       mov      .cur1r,edi
+       mov      .cur2r,edi
+       movzx    esi,word .col1g
+       shl      esi,ROUND
+       mov      .cur1g,esi
+       mov      .cur2g,esi
+       movzx    edx,word .col1b
+       shl      edx,ROUND
+       mov      .cur1b,edx
+       mov      .cur2b,edx
+       movzx    edi,word .tex_x1
+       shl      edi,ROUND
+       mov      .scan_x1,edi
+       mov      .scan_x2,edi
+       movzx    edx,word .tex_y1
+       shl      edx,ROUND
+       mov      .scan_y1,edx
+       mov      .scan_y2,edx
+       mov      cx,.y1
+       cmp      cx,.y2
+       jge      .loop1_end
+    .loop_1:
+    ;   push     eax ebx ebp
+       pushad
+       push     .tex_ptr
+       push     .scr_buff
+       push     .z_ptr
+       push     cx
+       push     .zz2
+       push     .scan_x2
+       push     dword .scan_y2
+       push     dword .cur2r
+       push     .cur2g
+       push     dword .cur2b
+       push     .zz1
+       push     .scan_x1
+       push     dword .scan_y1
+       push     dword .cur1r
+       push     .cur1g
+       push     dword .cur1b
+       sar      eax,ROUND
+       sar      ebx,ROUND
+       call     horizontal_tex_grd_line
+      ; pop      ebp ebx eax
+       popad
+if (Ext = MMX)|(Ext=SSE)
+       movq     mm0,.cur1b
+       movq     mm1,.cur1r
+       movq     mm2,.scan_y1
+       movq     mm3,.cur2b
+       movq     mm4,.cur2r
+       movq     mm5,.scan_y2
+       paddd    mm0,.dc13b
+       paddd    mm1,.dc13r
+       paddd    mm2,.tex_dy13
+       paddd    mm3,.dc12b
+       paddd    mm4,.dc12r
+       paddd    mm5,.tex_dy12
+       movq     .cur1b,mm0
+       movq     .cur1r,mm1
+       movq     .scan_y1,mm2
+       movq     .cur2b,mm3
+       movq     .cur2r,mm4
+       movq     .scan_y2,mm5
+end if
+if Ext >= SSE2
+       movups   xmm0,.cur1b
+       movups   xmm1,.dc13b
+       movups   xmm2,.cur2b
+       movups   xmm3,.dc12b
+       movq     mm2,.scan_y1
+       movq     mm5,.scan_y2
+       paddd    xmm0,xmm1
+       paddd    xmm2,xmm3
+       paddd    mm2,.tex_dy13
+       paddd    mm5,.tex_dy12
+       movq     .scan_y1,mm2
+       movq     .scan_y2,mm5
+       movups   .cur1b,xmm0
+       movups   .cur2b,xmm2
+end if
+if Ext = NON
+       mov      edx,.dc13b
+       add      .cur1b,edx
+       mov      esi,.dc13g
+       add      .cur1g,esi
+       mov      edi,.dc13r
+       add      .cur1r,edi
+       mov      edx,.dz13
+       add      .zz1,edx
+       mov      edx,.tex_dx13
+       add      .scan_x1,edx
+       mov      esi,.tex_dy13
+       add      .scan_y1,esi
+       mov      edi,.dc12b
+       add      .cur2b,edi
+       mov      esi,.dc12g
+       add      .cur2g,esi
+       mov      edx,.dc12r
+       add      .cur2r,edx
+       mov      edi,.tex_dx12
+       add      .scan_x2,edi
+       mov      esi,.tex_dy12
+       add      .scan_y2,esi
+       mov      edx,.dz12
+       add      .zz2,edx
+end if
+       add      eax,.dx13
+       add      ebx,.dx12
+       inc      cx
+       cmp      cx,.y2
+       jl      .loop_1
+ .loop1_end:
+       movzx    ecx,.y2
+       cmp      cx,.y3
+       jge      .loop2_end
+       movsx    ebx,.x2                    ; eax - cur x1
+       shl      ebx,ROUND                  ; ebx - cur x2
+       movsx    edx,word .z2
+       shl      edx,CATMULL_SHIFT
+;       mov      .zz1,edx
+       mov      .zz2,edx
+       movzx    edi,word .col2r
+       shl      edi,ROUND
+   ;    mov      .cur1r,edi
+       mov      .cur2r,edi
+       movzx    esi,word .col2g
+       shl      esi,ROUND
+   ;    mov      .cur1g,esi
+       mov      .cur2g,esi
+       movzx    edx,word .col2b
+       shl      edx,ROUND
+   ;    mov      .cur1b,edx
+       mov      .cur2b,edx
+       movzx    edi,word .tex_x2
+       shl      edi,ROUND
+ ;      mov      .scan_x1,edi
+       mov      .scan_x2,edi
+       movzx    edx,word .tex_y2
+       shl      edx,ROUND
+ ;      mov      .scan_y1,edx
+       mov      .scan_y2,edx
+  .loop_2:
+       pushad
+       push     .tex_ptr
+       push     .scr_buff
+       push     .z_ptr
+       push     cx
+       push     .zz2
+       push     .scan_x2
+       push     dword .scan_y2
+       push     dword .cur2r
+       push     .cur2g
+       push     dword .cur2b
+       push     .zz1
+       push     .scan_x1
+       push     dword .scan_y1
+       push     dword .cur1r
+       push     .cur1g
+       push     dword .cur1b
+       sar      eax,ROUND
+       sar      ebx,ROUND
+       call     horizontal_tex_grd_line
+       popad
+if (Ext = MMX)|(Ext=SSE)
+       movq     mm0,.cur1b
+       movq     mm1,.cur1r
+       movq     mm2,.scan_y1
+       movq     mm3,.cur2b
+       movq     mm4,.cur2r
+       movq     mm5,.scan_y2
+       paddd    mm0,.dc13b
+       paddd    mm1,.dc13r
+       paddd    mm2,.tex_dy13
+       paddd    mm3,.dc23b
+       paddd    mm4,.dc23r
+       paddd    mm5,.tex_dy23
+       movq     .cur1b,mm0
+       movq     .cur1r,mm1
+       movq     .scan_y1,mm2
+       movq     .cur2b,mm3
+       movq     .cur2r,mm4
+       movq     .scan_y2,mm5
+end if
+if Ext >= SSE2
+       movups   xmm0,.cur1b
+       movups   xmm1,.dc13b
+       movups   xmm2,.cur2b
+       movups   xmm3,.dc23b
+       movq     mm2,.scan_y1
+       movq     mm5,.scan_y2
+       paddd    xmm0,xmm1
+       paddd    xmm2,xmm3
+       paddd    mm2,.tex_dy13
+       paddd    mm5,.tex_dy23
+       movq     .scan_y1,mm2
+       movq     .scan_y2,mm5
+       movups   .cur1b,xmm0
+       movups   .cur2b,xmm2
+end if
+if Ext = NON
+       mov      edx,.dc13b
+       add      .cur1b,edx
+       mov      esi,.dc13g
+       add      .cur1g,esi
+       mov      edi,.dc13r
+       add      .cur1r,edi
+       mov      edx,.tex_dx13
+       add      .scan_x1,edx
+       mov      esi,.tex_dy13
+       add      .scan_y1,esi
+       mov      edx,.dz13
+       add      .zz1,edx
+       mov      edi,.dc23b
+       add      .cur2b,edi
+       mov      esi,.dc23g
+       add      .cur2g,esi
+       mov      edx,.dc23r
+       add      .cur2r,edx
+       mov      edi,.tex_dx23
+       add      .scan_x2,edi
+       mov      esi,.tex_dy23
+       add      .scan_y2,esi
+       mov      edx,.dz23
+       add      .zz2,edx
+end if
+       add      eax,.dx13
+       add      ebx,.dx23
+       inc      cx
+       cmp      cx,.y3
+       jl       .loop_2
+.loop2_end:
+       mov      esp,ebp
+ret 36
+horizontal_tex_grd_line:
+;in:
+; eax : x1, ebx : x2
+.tex_ptr  equ [ebp+62]
+.screen   equ [ebp+58]
+.z_buffer equ [ebp+54]
+.y        equ [ebp+52]
+.z2     equ [ebp+48]
+.tex_x2 equ [ebp+44]
+.tex_y2 equ [ebp+40]
+.r2     equ [ebp+36]
+.g2     equ [ebp+32]
+.b2     equ [ebp+28]
+.z1     equ [ebp+24]
+.tex_x1 equ [ebp+20]
+.tex_y1 equ [ebp+16]
+.r1     equ [ebp+12]
+.g1     equ [ebp+8]
+.b1     equ [ebp+4]
+.x1 equ  word[ebp-2]
+.x2 equ  word[ebp-4]
+.dz equ dword[ebp-8]
+.db equ dword[ebp-12]
+.dg equ dword[ebp-16]
+.dr equ dword[ebp-20]
+.dtex_x equ dword[ebp-24]
+.dtex_y equ dword[ebp-28]
+.c_ty equ [ebp-32]
+.c_tx equ [ebp-36]
+.cb  equ  [ebp-40]
+.cg  equ  [ebp-44]
+.cr  equ  [ebp-48]
+.t_col equ [ebp-52]
+.dtex_yM equ qword[ebp-28]
+.drM equ qword[ebp-20]
+.dbM equ qword[ebp-12]
+        mov     ebp,esp
+  ;      sub     esp,30
+        mov     cx,word .y
+        or      cx,cx
+        jl      .quit_l
+        cmp     cx,word[size_y_var]  ;SIZE_Y
+        jge     .quit_l
+        cmp     ax,bx
+        je      .quit_l
+        jl      @f
+        xchg    eax,ebx
+if Ext=NON
+        mov     ecx,dword .r1
+        xchg    ecx, .r2
+        mov     dword .r1, ecx
+        mov     ecx,dword .g1
+        xchg    ecx, .g2
+        mov     dword .g1, ecx
+        mov     ecx,dword .b1
+        xchg    ecx, .b2
+        mov     dword .b1, ecx
+        mov     ecx,dword .tex_x1
+        xchg    ecx, .tex_x2
+        mov     dword .tex_x1, ecx
+        mov     ecx,dword .tex_y1
+        xchg    ecx, .tex_y2
+        mov     dword .tex_y1, ecx
+        mov     ecx,dword .z1
+        xchg    ecx, .z2
+        mov     dword .z1, ecx
+end if
+if (Ext=MMX)
+        movq    mm0,.b1        ; b, g
+        movq    mm1,.b2
+        movq    .b1, mm1
+        movq    .b2, mm0
+        movq    mm2,.r1        ; r, y
+        movq    mm3,.r2
+        movq    .r1,mm3
+        movq    .r2,mm2
+        movq    mm4,.tex_x1    ; x, z
+        movq    mm5,.tex_x2
+        movq    .tex_x1,mm5
+        movq    .tex_x2,mm4
+end if
+if Ext>=SSE
+       movups  xmm0,.b1
+       movups  xmm1,.b2
+       movups  .b1,xmm1
+       movups  .b2,xmm0
+       movq    mm4,.tex_x1    ; x, z
+       movq    mm5,.tex_x2
+       movq    .tex_x1,mm5
+       movq    .tex_x2,mm4
+end if
+    @@:
+        or      bx,bx
+        jle     .quit_l
+        cmp     ax,word[size_x_var]  ;SIZE_X
+        jge     .quit_l
+        push    ax
+        push    bx
+        mov     eax,.z2            ; delta zone************
+        sub     eax,.z1
+        cdq
+        mov     bx,.x2
+        sub     bx,.x1
+        movsx   ebx,bx
+        idiv    ebx
+        push    eax  ; .dz
+        mov     eax,.b2
+        sub     eax,.b1
+        cdq
+        idiv    ebx
+        push    eax  ; .db
+        mov     eax,.g2
+        sub     eax,.g1
+        cdq
+        idiv    ebx
+        push    eax  ; .dg
+        mov     eax,.r2
+        sub     eax,.r1
+        cdq
+        idiv    ebx
+        push    eax  ; .dr
+        mov     eax,.tex_x2
+        sub     eax,.tex_x1
+        cdq
+        idiv    ebx
+        push    eax  ; .dtex_x
+        mov     eax,.tex_y2
+        sub     eax,.tex_y1
+        cdq
+        idiv    ebx
+        push    eax  ; .dtey_x
+        cmp     .x1,0
+        jg      @f
+        mov     eax,.dz     ; clipping
+        movsx   ebx,.x1
+        neg     ebx
+        imul    ebx
+        add     .z1,eax
+        mov     .x1,0
+        mov    eax,.dr
+        imul   ebx
+        add    .r1,eax
+;if  Ext=NON
+        mov    eax,.dg
+        imul   ebx
+        add    .g1,eax
+        mov    eax,.db
+        imul   ebx
+        add    .b1,eax
+        mov    eax,.dtex_x
+        imul   ebx
+        add    .tex_x1,eax
+        mov    eax,.dtex_y
+        imul   ebx
+        add    .tex_y1,eax
+   @@:
+        movsx     edx,word[size_x_var]  ;SIZE_X
+        cmp       .x2,dx
+        jl        @f
+        mov       .x2,dx
+    @@:
+; calc line addres begin in screen and Z buffer
+        movsx     eax,word .y
+        mul       edx
+        movsx     edx,.x1
+        add       eax,edx
+        mov       esi,eax
+        shl       esi,2
+        add       esi,.z_buffer
+        lea       eax,[eax*3]
+        mov       edi,.screen
+        add       edi,eax
+        mov       cx,.x2
+        sub       cx,.x1
+        movzx     ecx,cx
+; init current variables
+        push      dword .tex_y1
+;if  Ext=NON
+        push      dword .tex_x1
+        push      dword .b1
+        push      dword .g1
+        push      dword .r1
+if Ext>=MMX
+        movq      mm4,.cr  ; lo -> r,g
+        movq      mm6,.cb  ; hi -> b, tex_x
+        pxor      mm0,mm0
+end if
+        mov       ebx,.z1
+      .ddraw:
+        cmp       ebx,dword[esi]
+        jge       @f
+        mov       eax,.c_ty
+;  if ROUND<TEX_SHIFT
+;        shl        eax,TEX_SHIFT-ROUND
+;  end if
+;  if ROUND>TEX_SHIFT
+;        shr        eax,ROUND-TEX_SHIFT
+;  end if
+        shr       eax,ROUND
+        shl       Eax,TEX_SHIFT
+        mov       edx,.c_tx   ; calc texture pixel mem addres
+        shr       edx,ROUND
+        add       eax,edx
+        and       eax,TEXTURE_SIZE   ; cutting
+        lea       eax,[3*eax]
+        add       eax,.tex_ptr
+        mov       dword[esi],ebx
+if    Ext = NON
+        mov       eax,dword[eax]
+ ;      mov        .tex_col,eax
+        push      ax
+        shl       eax,8
+        pop       ax
+        mov       edx,.cr
+        sar       edx,ROUND
+        mul       dl         ; al*dl
+        shr       ax,8
+        stosb
+        ror       eax,16
+        push      ax
+        mov       edx,.cg
+        sar       edx,ROUND
+        mul       dl
+        shr       ax,8
+        stosb
+        pop       ax
+        shr       ax,8
+        mov       edx,.cb
+        sar       edx,ROUND
+        mul       dl
+        shr       ax,8
+        stosb
+        jmp       .no_skip
+else
+        movd      mm1,[eax]
+        punpcklbw mm1,mm0
+        movq      mm3,mm4  ;.cr     ; lo -> r,g
+        movq      mm5,mm6  ;.cb     ; lo -> b,tex_x
+        psrld     mm3,ROUND  ;
+        psrld     mm5,ROUND  ;
+        packssdw  mm3,mm5
+        pmullw    mm1,mm3
+        psrlw     mm1,8
+        packuswb  mm1,mm0
+        movd      [edi],mm1
+end if
+       mov        dword[esi],ebx
+if Ext = NON
+       jmp        .no_skip
+end if
+     @@:
+       add        edi,3
+     .no_skip:
+       add        esi,4
+       add        ebx,.dz
+       mov        eax,.dtex_x
+       add        .c_tx, eax
+       mov        edx,.dtex_y
+       add        .c_ty, edx
+if Ext=NON
+       mov        eax,.dr
+       add        .cr,eax
+       mov        edx,.dg
+       add        .cg,edx
+       mov        eax,.db
+       add        .cb,eax
+else
+        paddd      mm4,.drM
+        paddd      mm6,.dbM
+;;        paddd      mm7,.dtex_y      ; mm4 - b, g
+;;        movq       .c_tx,mm7
+                                    ; mm6 - r, x
+end if                              ; mm7 - y, x
+       dec      ecx
+       jnz      .ddraw
+  .quit_l:
+        mov     esp,ebp
+ret 42+20  ; horizontal line

 /programs/demos/view3ds/history.txt
 ,11 → 1,16
+View3ds 0.073 - may 2021
+. I introduced procedure for searching nonredundand edges.
+. Writing some info about object: vertices, triangles unique edges
+    count.
+-----------------------------------------------------------------------------------
 View3ds 0.072 - march 2021
 . New displaying model - texturing with bilinear filtering and transparency
-    simultanusly. Note that filtering is done only inside polygon. To better
+    simultanusly. Note that filtering is done only inside polygon. To better
     quality of image there is a need to use floats coordinates of texture to pass
     as arguments to single triangle rendering proc.
 . Optimizations.
-. SSE3 version runs correct on SSE2 cpus, but real phong, glass and
+. SSE3 version runs correct on SSE2 cpus, but real phong, glass and
     transparented texturing with filtering rendering models are disabled.
 -----------------------------------------------------------------------------------

 /programs/demos/view3ds/readme.txt
 ,20 → 1,16
-View3ds 0.073 - tiny viewer to .3ds and .asc files with several graphics
+View3ds 0.074 - tiny viewer to .3ds and .asc files with several graphics
                 effects implementation.
 What's new?
-. I introduced procedure for searching nonredundand edges.
-. Writing some info about object: vertices, triangles unique edges
-    count.
+. Fixed emboss bug in grd lines displaying model.
+. Grd line exceedes screen problem fix.
+. New rendering model - ray casted shadows and appropiate button to
+   set 'on' this option. Note that is non real time model, especially when
+   complex object is computed. I took effort to introduce accelerating
+   structure - AABB (Axis Aligned Bounding Boxes).. but it is disabled
+   for now - seems to work incorrect(slow).
-. New displaying model - texturing with bilinear filtering and transparency
-    simultanusly. Note that filtering is done only inside polygon. To better
-    quality of image there is a need to use floats coordinates of texture to pass
-    as arguments to single triangle rendering proc.
-. Optimizations.
-. SSE3 version runs correct on SSE2 cpus, but real phong, glass and
-    transparented texturing with filtering rendering models are disabled.
 Buttons description:
 .  rotary: choosing rotary axle: x, y, x+y.
 .  shd. model: choosing shading model: flat, grd (smooth), env (spherical
 ,20 → 22,20
     ptex (real Phong + texturing + transparency).
 .  speed: idle, full.
 ,5. zoom in, out: no comment.
-.  catmull: disabled
+.  ray shadow: calc ray casted shadows.
 .  culling: backface culling on/ off.
 .  rand. light: Randomize 3 unlinear lights( so called Phong's illumination).
 .  Blur: blur N times; N=0,1,2,3,4,5
 .11,12,13. loseless operations (rotary 90, 180 degrees).
-. emboss: Do emboss effect( flat bumps ), use 'bumps deep' button to do edges
+. emboss: Do emboss effect( flat bumps ), use 'bumps deep' button to do edges
      more deep.
 . fire: do motion blur ( looks like fire ).
-. move: changes meaning x,y,z +/- buttons  ->  obj: moving object, camr: moving
+. move: changes meaning x,y,z +/- buttons  ->  obj: moving object, camr: moving
       camera, wave: x,y +/- increase, decrease wave effect frequency and amplitude.
 . generate: Generates some objects: node, Thorn Crown, heart...
 . bumps: random, according to texture.
 . bumps deep -> create bumps deeper or lighter.
-. re-map tex -> re-map texture and bump map coordinates, to change spherical
+. re-map tex -> re-map texture and bump map coordinates, to change spherical
      mapping  around axle use 'xchg' and 'mirror' buttons, then press 're-map tex' button.
 . bright + -> increase picture brightness.
 . bright - -> decrease picture brightness.
 ,8 → 42,8
 . wav effect -> do effect based sine function.
 . editor -> setting editing option. If is "on" then red bars are draw according to each
    vertex, Pressing  and moving left mouse button (cursor must be on handler)- change
-   vertex position. If left mouse button is released apply current position. You may also
+   vertex position. If left mouse button is released apply current position. You may also
    decrease whole handlers count by enable culling (using appropriate button) - some
    back handlers become hidden.
-                         Maciej Guba             V 2021
+                         Maciej Guba             IX 2021

 /programs/demos/view3ds/tex_cat.inc
 ,611 → 1,611
-;TEX_X = 512
-;TEX_Y = 512
-;ROUND equ 8
-;SIZE_X = 512
-;SIZE_Y = 512
-;TEX_SHIFT = 9
-CATMULL_SHIFT equ 8
-;------------------------------------------------------------------------
-;- Procedure drawing textured triangle using Catmull Z-buffer algorithm -
-;------------------------------------------------------------------------
-tex_triangle_z:
-;----------in - eax - x1 shl 16 + y1
-;-------------- ebx - x2 shl 16 + y2
-;---------------ecx - x3 shl 16 + y3
-;---------------edx - pointer to Z-buffer
-;---------------esi - pointer to texture buffer
-;---------------edi - pointer to screen buffer
-;-------------stack - texture coordinates
-;------------------ - z coordinates
-.tex_x1 equ  ebp+4
-.tex_y1 equ  ebp+6
-.tex_x2 equ  ebp+8
-.tex_y2 equ  ebp+10
-.tex_x3 equ  ebp+12
-.tex_y3 equ  ebp+14
-.z1     equ  word[ebp+16]
-.z2     equ  word[ebp+18]
-.z3     equ  word[ebp+20]
-.tex_ptr equ dword[ebp-4]        ; pointer to texture
-.z_ptr   equ dword[ebp-8]        ; pointer to z-buffer
-.x1      equ word[ebp-10]
-.y1      equ word[ebp-12]
-.x2      equ word[ebp-14]
-.y2      equ word[ebp-16]
-.x3      equ word[ebp-18]
-.y3      equ word[ebp-20]
-.dx12     equ dword[ebp-24]
-.tex_dx12 equ dword[ebp-28]
-.tex_dy12 equ dword[ebp-32]
-.dz12     equ dword[ebp-36]
-.dx13     equ dword[ebp-40]
-.tex_dx13 equ dword[ebp-44]
-.tex_dy13 equ dword[ebp-48]
-.dz13     equ dword[ebp-52]
-.dx23     equ dword[ebp-56]
-.tex_dx23 equ dword[ebp-60]
-.tex_dy23 equ dword[ebp-64]
-.dz23     equ dword[ebp-68]
-.scan_x1  equ dword[ebp-72]
-.scan_x2  equ dword[ebp-76]
-.scan_y1  equ dword[ebp-80]
-.scan_y2  equ dword[ebp-84]
-.cz1      equ dword[ebp-88]
-.cz2      equ dword[ebp-92]
-        mov     ebp,esp
-        push    esi               ; store memory pointers
-        push    edx
-.tt_sort3:
-        cmp     ax,bx                     ;sort all parameters
-        jle     .tt_sort1
-        xchg    eax,ebx
-        mov     edx,dword [.tex_x1]
-        xchg    edx,dword [.tex_x2]
-        mov     dword[.tex_x1],edx
-        mov     dx,.z1
-        xchg    dx,.z2
-        mov     .z1,dx
-.tt_sort1:
-        cmp     bx,cx
-        jle     .tt_sort2
-        xchg    ebx,ecx
-        mov     edx,dword [.tex_x2]
-        xchg    edx,dword [.tex_x3]
-        mov     dword [.tex_x2],edx
-        mov     dx,.z2
-        xchg    dx,.z3
-        mov     .z2,dx
-        jmp     .tt_sort3
-.tt_sort2:
-        push    eax     ; and store to user friendly variables
-        push    ebx
-        push    ecx
-        mov      edx,80008000h  ; eax,ebx,ecx are ANDd together into edx which means that
-        and      edx,ebx        ; if *all* of them are negative a sign flag is raised
-        and      edx,ecx
-        and      edx,eax
-        test     edx,80008000h  ; Check both X&Y at once
-        jne      .tt_loop2_end
-    ;    cmp      ax,SIZE_Y
-    ;    jl       @f
-    ;    cmp      bx,SIZE_Y
-    ;    jl       @f
-    ;    cmp      cx,SIZE_Y
-    ;    jl       @f
-        ror      eax,16
-        ror      ebx,16
-        ror      ecx,16
-    ;    cmp      ax,SIZE_X
-    ;    jl       @f
-    ;    cmp      bx,SIZE_X
-    ;    jl       @f
-    ;    cmp      cx,SIZE_X
-    ;    jl       @f
-    ;    jmp      .tt_loop2_end
-     @@:
-        mov      eax,dword[.tex_x1] ; texture coords must be in [0..TEX_X(Y)]
-        mov      ebx,dword[.tex_x2]
-        mov      ecx,dword[.tex_x3]
-        mov      edx,eax
-        or       edx,ebx
-        or       edx,ecx
-        test     edx,80008000h
-        jne      .tt_loop2_end
-        cmp      ax,TEX_X
-        jge      .tt_loop2_end
-        cmp      bx,TEX_X
-        jge      .tt_loop2_end
-        cmp      cx,TEX_X
-        jge      .tt_loop2_end
-        ror      eax,16
-        ror      ebx,16
-        ror      ecx,16
-        cmp      ax,TEX_Y
-        jge      .tt_loop2_end
-        cmp      bx,TEX_Y
-        jge      .tt_loop2_end
-        cmp      cx,TEX_Y
-        jge      .tt_loop2_end
-        movsx    ebx,.y2    ; calc delta
-        sub      bx,.y1
-        jnz      .tt_dx12_make
-        xor      edx,edx
-        mov      ecx,4
-    @@:
-        push     edx
-        loop     @b
-        jmp      .tt_dx12_done
-    .tt_dx12_make:
-        mov      ax,.x2
-        sub      ax,.x1
-        cwde
-        shl      eax,ROUND
-        cdq
-        idiv     ebx
-       ; mov      .dx12,eax                      ; dx12 = (x2-x1)/(y2-y1)
-        push     eax
-        mov      ax,word[.tex_x2]
-        sub      ax,word[.tex_x1]
-        cwde
-        shl      eax,ROUND
-        cdq
-        idiv     ebx
-       ; mov     [.tex_dx12],eax               ; tex_dx12 = (tex_x2-tex_x1)/(y2-y1)
-        push     eax
-        mov      ax,word[.tex_y2]
-        sub      ax,word[.tex_y1]
-        cwde
-        shl      eax,ROUND
-        cdq
-        idiv     ebx
-     ;   mov    [.tex_dy12],eax              ; tex_dy12 = (tex_y2-tex_y1)/(y2-y1)
-        push     eax
-        mov      ax,.z2
-        sub      ax,.z1
-        cwde
-        shl      eax,CATMULL_SHIFT
-        cdq
-        idiv     ebx
-        push     eax
-   .tt_dx12_done:
-        movsx    ebx,.y3    ; calc delta
-        sub      bx,.y1
-        jnz      .tt_dx13_make
-        xor      edx,edx
-        mov      ecx,4
-    @@:
-        push     edx
-        loop     @b
-        jmp      .tt_dx13_done
-    .tt_dx13_make:
-        mov      ax,.x3
-        sub      ax,.x1
-        cwde
-        shl      eax,ROUND
-        cdq
-        idiv     ebx
-       ; mov      .dx12,eax                      ; dx13 = (x3-x1)/(y3-y1)
-        push      eax
-        mov      ax,word[.tex_x3]
-        sub      ax,word[.tex_x1]
-        cwde
-        shl      eax,ROUND
-        cdq
-        idiv     ebx
-       ; mov     [.tex_dx12],eax               ; tex_dx13 = (tex_x3-tex_x1)/(y3-y1)
-        push     eax
-        mov      ax,word[.tex_y3]
-        sub      ax,word[.tex_y1]
-        cwde
-        shl      eax,ROUND
-        cdq
-        idiv     ebx
-     ;   mov    [.tex_dy12],eax              ; tex_dy13 = (tex_y3-tex_y1)/(y3-y1)
-        push     eax
-        mov      ax,.z3
-        sub      ax,.z1
-        cwde
-        shl      eax,CATMULL_SHIFT
-        cdq
-        idiv     ebx
-        push     eax
-   .tt_dx13_done:
-        mov      bx,.y3    ; calc delta
-        sub      bx,.y2
-        jnz      .tt_dx23_make
-        xor      edx,edx
-        mov      ecx,4
-    @@:
-        push     edx
-        loop     @b
-        jmp      .tt_dx23_done
-    .tt_dx23_make:
-        mov      ax,.x3
-        sub      ax,.x2
-        cwde
-        shl      eax,ROUND
-        cdq
-        movzx    ebx,bx
-        idiv     ebx
-       ; mov      .dx23,eax                      ; dx23 = (x3-x2)/(y3-y2)
-        push      eax
-        mov      ax,word[.tex_x3]
-        sub      ax,word[.tex_x2]
-        cwde
-        shl      eax,ROUND
-        cdq
-        idiv     ebx
-       ; mov     [.tex_dx23],eax               ; tex_dx23 = (tex_x3-tex_x2)/(y3-y2)
-        push     eax
-        mov      ax,word[.tex_y3]
-        sub      ax,word[.tex_y2]
-        cwde
-        shl      eax,ROUND
-        cdq
-        idiv     ebx
-     ;   mov    [.tex_dy23],eax              ; tex_dy23 = (tex_y3-tex_y2)/(y3-y2)
-        push     eax
-        mov      ax,.z3
-        sub      ax,.z2
-        cwde
-        shl      eax,CATMULL_SHIFT
-        cdq
-        idiv     ebx
-        push     eax
-   .tt_dx23_done:
-        movsx    eax,.x1     ;eax - cur x1
-        shl      eax,ROUND   ;ebx - cur x2
-        mov      ebx,eax
-        movsx    edx, word[.tex_x1]
-        shl      edx,ROUND
-      ;  mov [.scan_x1],edx
-      ;  mov [.scan_x2],edx
-        push     edx
-        push     edx
-        movsx    edx, word[.tex_y1]
-        shl      edx,ROUND
-       ; mov [.scan_y1],edx
-       ; mov [.scan_y2],edx
-        push     edx
-        push     edx
-        movsx    edx,.z1
-        shl      edx,CATMULL_SHIFT
-        push     edx
-        push     edx
-        mov      cx,.y1
-        cmp      cx,.y2
-        jge      .tt_loop1_end
-   .tt_loop1:
-        pushad
-        push    .z_ptr
-        push    .cz1      ; z coords shifted shl catmull_shift
-        push    .cz2
-        push    .scan_y2
-        push    .scan_x2
-        push    .scan_y1
-        push    .scan_x1
-        push    esi               ;[.tex_ptr]
-        push    cx
-        sar     ebx,ROUND
-        push    bx
-        sar     eax,ROUND
-        push    ax
-        call    textured_line_z
-        popad
-        mov     edx,.dz13
-        add     .cz1,edx
-        mov     edx,.dz12
-        add     .cz2,edx
-        mov     edx, .tex_dx13
-        add     .scan_x1, edx
-        mov     edx, .tex_dx12
-        add     .scan_x2, edx
-        mov     edx, .tex_dy13
-        add     .scan_y1, edx
-        mov     edx, .tex_dy12
-        add     .scan_y2, edx
-        add     eax, .dx13
-        add     ebx, .dx12
-        inc     cx
-        cmp     cx,.y2
-        jl      .tt_loop1
-  .tt_loop1_end:
-        mov     cx,.y2
-        cmp     cx,.y3
-        jge     .tt_loop2_end
-        movsx   ebx,.x2
-        shl     ebx,ROUND
-        movsx   edx,.z2
-        shl     edx,CATMULL_SHIFT
-        mov     .cz2,edx
-        movzx   edx, word [.tex_x2]
-        shl     edx,ROUND
-        mov     .scan_x2,edx
-        movzx   edx, word[.tex_y2]
-        shl     edx,ROUND
-        mov     .scan_y2,edx
-.tt_loop2:
-        pushad
-        push    .z_ptr
-        push    .cz1      ; z coords shifted shl catmull_shift
-        push    .cz2
-        push    .scan_y2
-        push    .scan_x2
-        push    .scan_y1
-        push    .scan_x1
-        push    esi               ;[.tex_ptr]
-        push    cx
-        sar     ebx,ROUND
-        push    bx
-        sar     eax,ROUND
-        push    ax
-        call    textured_line_z
-        popad
-        mov     edx,.dz13
-        add     .cz1,edx
-        mov     edx,.dz23
-        add     .cz2,edx
-        mov     edx, .tex_dx13
-        add     .scan_x1, edx
-        mov     edx, .tex_dx23
-        add     .scan_x2, edx
-        mov     edx, .tex_dy13
-        add     .scan_y1, edx
-        mov     edx, .tex_dy23
-        add     .scan_y2, edx
-        add     eax, .dx13
-        add     ebx, .dx23
-        inc     cx
-        cmp     cx,.y3
-        jl      .tt_loop2
-.tt_loop2_end:
-.tt_end:
-        mov esp,ebp
-ret 18
-textured_line_z:
-;-----in -edi screen buffer pointer
-;------------ stack:
-  .x1 equ word [ebp+4]
-  .x2 equ word [ebp+6]
-  .y  equ word [ebp+8]
-  .tex_ptr equ dword [ebp+10]
-  .tex_x1  equ ebp+14
-  .tex_y1  equ ebp+18
-  .tex_x2  equ ebp+22
-  .tex_y2  equ ebp+26
-  .z2      equ dword [ebp+30]     ;z1, z2  coords shifted shl CATMULL_SHIFT
-  .z1      equ dword [ebp+34]
-  .z_ptr   equ dword [ebp+38]
-  .tex_dy  equ dword [ebp-4]
-  .tex_dx  equ dword [ebp-8]
-  .dz      equ dword [ebp-12]
-  .cz      equ dword [ebp-16]
-  .c_tex_x equ dword [ebp-20]  ; current tex x
-  .m_sft1 equ ebp-28
-  .m_sft2 equ ebp-32
-;  .c_tex_xM equ ebp+14
-  .tex_dxM  equ ebp-8
-        mov     ebp,esp
-        mov     ax,.y
-        or      ax,ax
-        jl      .tl_quit
-        mov     bx,[size_y_var]
-        dec     bx
-        cmp     ax,bx  ;SIZE_Y
-        jge     .tl_quit
-        mov     ax,.x1
-        cmp     ax,.x2
-        je      .tl_quit
-        jl      .tl_ok
-        xchg    ax,.x2    ; sort params
-        mov     .x1,ax
-if Ext >= MMX
-        movq    mm0,[.tex_x1]
-        movq    mm1,[.tex_x2]
-        movq    [.tex_x2],mm0
-        movq    [.tex_x1],mm1
-else
-        mov     eax,dword[.tex_x1]
-        xchg    eax,dword[.tex_x2]
-        mov     dword[.tex_x1],eax
-        mov     eax,dword[.tex_y1]
-        xchg    eax,dword[.tex_y2]
-        mov     dword[.tex_y1],eax
-end if
-        mov     eax,.z1
-        xchg    eax,.z2
-        mov     .z1,eax
-     .tl_ok:
-        mov     cx,[size_x_var]
-        dec     cx
-        cmp     .x1,cx   ;SIZE_X
-        jge     .tl_quit
-        cmp     .x2,0
-        jle     .tl_quit
-        mov     bx,.x2
-        sub     bx,.x1
-        movsx   ebx,bx
-        mov     eax,dword[.tex_y2]       ; calc .dty
-        sub     eax,dword[.tex_y1]
-        cdq
-        idiv    ebx
-        push    eax
-        mov     eax,dword[.tex_x2]       ; calc .dtx
-        sub     eax,dword[.tex_x1]
-        cdq
-        idiv    ebx
-        push    eax
-        mov     eax,.z2       ; calc .dz
-        sub     eax,.z1
-        cdq
-        idiv    ebx
-        push    eax
-        cmp    .x1,0          ; clipping
-        jg     @f
-        movsx   ebx,.x1
-        neg     ebx
-        imul    ebx           ; eax = .dz * abs(.x1)
-        add     .z1,eax
-        mov     .x1,0
-        mov     eax,.tex_dy
-        imul    ebx
-        add     dword[.tex_y1],eax
-        mov     eax,.tex_dx
-        imul    ebx
-        add     dword[.tex_x1],eax
-      @@:
-        cmp     .x2,cx ;SIZE_X
-        jl      @f
-        mov     .x2,cx  ;SIZE_X
-      @@:
-        movsx   ebx,.y        ; calc mem begin in buffers
-        movzx   eax,word[size_x_var]  ;SIZE_X
-        mul     ebx
-        movsx   ebx,.x1
-        add     eax,ebx
-        mov     ebx,eax
-        lea     eax,[eax*3]
-        add     edi,eax               ; edi - scr buff
-        shl     ebx,2
-        add     .z_ptr,ebx            ; z buffer pointer
-        mov     cx,.x2
-        sub     cx,.x1
-        movzx   ecx,cx
-;if Ext >= MMX
-;        movq    mm0,[.tex_x1]
-;        movq    mm4,mm0
-;        movq    mm1,qword[.tex_dxM]
-;        mov     ebx,.z1
-;        mov     eax,.dz
-;else
-        mov     eax,dword[.tex_x1]
-        mov     ebx,dword[.tex_y1]
-        push    .z1           ; .cz
-        push    eax           ;.c_tex_x
-;end if
-        mov     edx,.z_ptr
-     .tl_loop:
-;if Ext >= MMX
-;        cmp     ebx,[edx]                       ;  ebx - current z
-;        jge     @f
-;        movq    mm2,mm0
-;        psrad   mm2,ROUND
-;        movq    mm3,mm2
-;        psrlq   mm2,32-TEX_SHIFT
-;        paddd   mm3,mm2
-;        movd    esi,mm3
-;        mov     dword[edx],ebx    ; renew z buffer
-;else
-                                                        ;  eax - temp
-        mov     eax,.cz                         ;  ebx - cur tex y shl ROUND
-        cmp     eax,[edx]                       ;  ecx - l.lenght
-        jge     @f         ; ebx - cur tex_y ;  edx - temp
-        mov     esi,ebx                 ;  edi - scr buff
-        sar     esi,ROUND               ;  esi - tex_ptr temp
-        shl     esi,TEX_SHIFT           ;  .z_ptr - cur pointer to z buff
-        mov     eax,.c_tex_x            ;  .cz - cur z coord shl CATMULL_SHIFT
-        sar     eax,ROUND
-        add     esi,eax
-        mov     eax,.cz
-        mov     dword[edx],eax    ; renew z buffer
-;end if
-        and     esi,TEXTURE_SIZE
-        lea     esi,[esi*3]
-        add     esi,.tex_ptr
-        movsd
-        dec     edi
-        jmp     .no_skip
-      @@:
-        add     edi,3
-     .no_skip:
-        add     edx,4
-;if Ext >= MMX
-;        add     ebx,eax
-;        paddd   mm0,mm1
-;else
-        mov     eax,.dz
-        add     .cz,eax
-        mov     eax,.tex_dx
-        add     .c_tex_x,eax
-        add     ebx,.tex_dy
-;end if
-        loop    .tl_loop
- .tl_quit:
-        mov     esp,ebp
-ret 30+8
+;TEX_X = 512
+;TEX_Y = 512
+;ROUND equ 8
+;SIZE_X = 512
+;SIZE_Y = 512
+;TEX_SHIFT = 9
+CATMULL_SHIFT equ 8
+;------------------------------------------------------------------------
+;- Procedure drawing textured triangle using Catmull Z-buffer algorithm -
+;------------------------------------------------------------------------
+tex_triangle_z:
+;----------in - eax - x1 shl 16 + y1
+;-------------- ebx - x2 shl 16 + y2
+;---------------ecx - x3 shl 16 + y3
+;---------------edx - pointer to Z-buffer
+;---------------esi - pointer to texture buffer
+;---------------edi - pointer to screen buffer
+;-------------stack - texture coordinates
+;------------------ - z coordinates
+.tex_x1 equ  ebp+4
+.tex_y1 equ  ebp+6
+.tex_x2 equ  ebp+8
+.tex_y2 equ  ebp+10
+.tex_x3 equ  ebp+12
+.tex_y3 equ  ebp+14
+.z1     equ  word[ebp+16]
+.z2     equ  word[ebp+18]
+.z3     equ  word[ebp+20]
+.tex_ptr equ dword[ebp-4]        ; pointer to texture
+.z_ptr   equ dword[ebp-8]        ; pointer to z-buffer
+.x1      equ word[ebp-10]
+.y1      equ word[ebp-12]
+.x2      equ word[ebp-14]
+.y2      equ word[ebp-16]
+.x3      equ word[ebp-18]
+.y3      equ word[ebp-20]
+.dx12     equ dword[ebp-24]
+.tex_dx12 equ dword[ebp-28]
+.tex_dy12 equ dword[ebp-32]
+.dz12     equ dword[ebp-36]
+.dx13     equ dword[ebp-40]
+.tex_dx13 equ dword[ebp-44]
+.tex_dy13 equ dword[ebp-48]
+.dz13     equ dword[ebp-52]
+.dx23     equ dword[ebp-56]
+.tex_dx23 equ dword[ebp-60]
+.tex_dy23 equ dword[ebp-64]
+.dz23     equ dword[ebp-68]
+.scan_x1  equ dword[ebp-72]
+.scan_x2  equ dword[ebp-76]
+.scan_y1  equ dword[ebp-80]
+.scan_y2  equ dword[ebp-84]
+.cz1      equ dword[ebp-88]
+.cz2      equ dword[ebp-92]
+        mov     ebp,esp
+        push    esi               ; store memory pointers
+        push    edx
+.tt_sort3:
+        cmp     ax,bx                     ;sort all parameters
+        jle     .tt_sort1
+        xchg    eax,ebx
+        mov     edx,dword [.tex_x1]
+        xchg    edx,dword [.tex_x2]
+        mov     dword[.tex_x1],edx
+        mov     dx,.z1
+        xchg    dx,.z2
+        mov     .z1,dx
+.tt_sort1:
+        cmp     bx,cx
+        jle     .tt_sort2
+        xchg    ebx,ecx
+        mov     edx,dword [.tex_x2]
+        xchg    edx,dword [.tex_x3]
+        mov     dword [.tex_x2],edx
+        mov     dx,.z2
+        xchg    dx,.z3
+        mov     .z2,dx
+        jmp     .tt_sort3
+.tt_sort2:
+        push    eax     ; and store to user friendly variables
+        push    ebx
+        push    ecx
+        mov      edx,80008000h  ; eax,ebx,ecx are ANDd together into edx which means that
+        and      edx,ebx        ; if *all* of them are negative a sign flag is raised
+        and      edx,ecx
+        and      edx,eax
+        test     edx,80008000h  ; Check both X&Y at once
+        jne      .tt_loop2_end
+    ;    cmp      ax,SIZE_Y
+    ;    jl       @f
+    ;    cmp      bx,SIZE_Y
+    ;    jl       @f
+    ;    cmp      cx,SIZE_Y
+    ;    jl       @f
+        ror      eax,16
+        ror      ebx,16
+        ror      ecx,16
+    ;    cmp      ax,SIZE_X
+    ;    jl       @f
+    ;    cmp      bx,SIZE_X
+    ;    jl       @f
+    ;    cmp      cx,SIZE_X
+    ;    jl       @f
+    ;    jmp      .tt_loop2_end
+     @@:
+        mov      eax,dword[.tex_x1] ; texture coords must be in [0..TEX_X(Y)]
+        mov      ebx,dword[.tex_x2]
+        mov      ecx,dword[.tex_x3]
+        mov      edx,eax
+        or       edx,ebx
+        or       edx,ecx
+        test     edx,80008000h
+        jne      .tt_loop2_end
+        cmp      ax,TEX_X
+        jge      .tt_loop2_end
+        cmp      bx,TEX_X
+        jge      .tt_loop2_end
+        cmp      cx,TEX_X
+        jge      .tt_loop2_end
+        ror      eax,16
+        ror      ebx,16
+        ror      ecx,16
+        cmp      ax,TEX_Y
+        jge      .tt_loop2_end
+        cmp      bx,TEX_Y
+        jge      .tt_loop2_end
+        cmp      cx,TEX_Y
+        jge      .tt_loop2_end
+        movsx    ebx,.y2    ; calc delta
+        sub      bx,.y1
+        jnz      .tt_dx12_make
+        xor      edx,edx
+        mov      ecx,4
+    @@:
+        push     edx
+        loop     @b
+        jmp      .tt_dx12_done
+    .tt_dx12_make:
+        mov      ax,.x2
+        sub      ax,.x1
+        cwde
+        shl      eax,ROUND
+        cdq
+        idiv     ebx
+       ; mov      .dx12,eax                      ; dx12 = (x2-x1)/(y2-y1)
+        push     eax
+        mov      ax,word[.tex_x2]
+        sub      ax,word[.tex_x1]
+        cwde
+        shl      eax,ROUND
+        cdq
+        idiv     ebx
+       ; mov     [.tex_dx12],eax               ; tex_dx12 = (tex_x2-tex_x1)/(y2-y1)
+        push     eax
+        mov      ax,word[.tex_y2]
+        sub      ax,word[.tex_y1]
+        cwde
+        shl      eax,ROUND
+        cdq
+        idiv     ebx
+     ;   mov    [.tex_dy12],eax              ; tex_dy12 = (tex_y2-tex_y1)/(y2-y1)
+        push     eax
+        mov      ax,.z2
+        sub      ax,.z1
+        cwde
+        shl      eax,CATMULL_SHIFT
+        cdq
+        idiv     ebx
+        push     eax
+   .tt_dx12_done:
+        movsx    ebx,.y3    ; calc delta
+        sub      bx,.y1
+        jnz      .tt_dx13_make
+        xor      edx,edx
+        mov      ecx,4
+    @@:
+        push     edx
+        loop     @b
+        jmp      .tt_dx13_done
+    .tt_dx13_make:
+        mov      ax,.x3
+        sub      ax,.x1
+        cwde
+        shl      eax,ROUND
+        cdq
+        idiv     ebx
+       ; mov      .dx12,eax                      ; dx13 = (x3-x1)/(y3-y1)
+        push      eax
+        mov      ax,word[.tex_x3]
+        sub      ax,word[.tex_x1]
+        cwde
+        shl      eax,ROUND
+        cdq
+        idiv     ebx
+       ; mov     [.tex_dx12],eax               ; tex_dx13 = (tex_x3-tex_x1)/(y3-y1)
+        push     eax
+        mov      ax,word[.tex_y3]
+        sub      ax,word[.tex_y1]
+        cwde
+        shl      eax,ROUND
+        cdq
+        idiv     ebx
+     ;   mov    [.tex_dy12],eax              ; tex_dy13 = (tex_y3-tex_y1)/(y3-y1)
+        push     eax
+        mov      ax,.z3
+        sub      ax,.z1
+        cwde
+        shl      eax,CATMULL_SHIFT
+        cdq
+        idiv     ebx
+        push     eax
+   .tt_dx13_done:
+        mov      bx,.y3    ; calc delta
+        sub      bx,.y2
+        jnz      .tt_dx23_make
+        xor      edx,edx
+        mov      ecx,4
+    @@:
+        push     edx
+        loop     @b
+        jmp      .tt_dx23_done
+    .tt_dx23_make:
+        mov      ax,.x3
+        sub      ax,.x2
+        cwde
+        shl      eax,ROUND
+        cdq
+        movzx    ebx,bx
+        idiv     ebx
+       ; mov      .dx23,eax                      ; dx23 = (x3-x2)/(y3-y2)
+        push      eax
+        mov      ax,word[.tex_x3]
+        sub      ax,word[.tex_x2]
+        cwde
+        shl      eax,ROUND
+        cdq
+        idiv     ebx
+       ; mov     [.tex_dx23],eax               ; tex_dx23 = (tex_x3-tex_x2)/(y3-y2)
+        push     eax
+        mov      ax,word[.tex_y3]
+        sub      ax,word[.tex_y2]
+        cwde
+        shl      eax,ROUND
+        cdq
+        idiv     ebx
+     ;   mov    [.tex_dy23],eax              ; tex_dy23 = (tex_y3-tex_y2)/(y3-y2)
+        push     eax
+        mov      ax,.z3
+        sub      ax,.z2
+        cwde
+        shl      eax,CATMULL_SHIFT
+        cdq
+        idiv     ebx
+        push     eax
+   .tt_dx23_done:
+        movsx    eax,.x1     ;eax - cur x1
+        shl      eax,ROUND   ;ebx - cur x2
+        mov      ebx,eax
+        movsx    edx, word[.tex_x1]
+        shl      edx,ROUND
+      ;  mov [.scan_x1],edx
+      ;  mov [.scan_x2],edx
+        push     edx
+        push     edx
+        movsx    edx, word[.tex_y1]
+        shl      edx,ROUND
+       ; mov [.scan_y1],edx
+       ; mov [.scan_y2],edx
+        push     edx
+        push     edx
+        movsx    edx,.z1
+        shl      edx,CATMULL_SHIFT
+        push     edx
+        push     edx
+        mov      cx,.y1
+        cmp      cx,.y2
+        jge      .tt_loop1_end
+   .tt_loop1:
+        pushad
+        push    .z_ptr
+        push    .cz1      ; z coords shifted shl catmull_shift
+        push    .cz2
+        push    .scan_y2
+        push    .scan_x2
+        push    .scan_y1
+        push    .scan_x1
+        push    esi               ;[.tex_ptr]
+        push    cx
+        sar     ebx,ROUND
+        push    bx
+        sar     eax,ROUND
+        push    ax
+        call    textured_line_z
+        popad
+        mov     edx,.dz13
+        add     .cz1,edx
+        mov     edx,.dz12
+        add     .cz2,edx
+        mov     edx, .tex_dx13
+        add     .scan_x1, edx
+        mov     edx, .tex_dx12
+        add     .scan_x2, edx
+        mov     edx, .tex_dy13
+        add     .scan_y1, edx
+        mov     edx, .tex_dy12
+        add     .scan_y2, edx
+        add     eax, .dx13
+        add     ebx, .dx12
+        inc     cx
+        cmp     cx,.y2
+        jl      .tt_loop1
+  .tt_loop1_end:
+        mov     cx,.y2
+        cmp     cx,.y3
+        jge     .tt_loop2_end
+        movsx   ebx,.x2
+        shl     ebx,ROUND
+        movsx   edx,.z2
+        shl     edx,CATMULL_SHIFT
+        mov     .cz2,edx
+        movzx   edx, word [.tex_x2]
+        shl     edx,ROUND
+        mov     .scan_x2,edx
+        movzx   edx, word[.tex_y2]
+        shl     edx,ROUND
+        mov     .scan_y2,edx
+.tt_loop2:
+        pushad
+        push    .z_ptr
+        push    .cz1      ; z coords shifted shl catmull_shift
+        push    .cz2
+        push    .scan_y2
+        push    .scan_x2
+        push    .scan_y1
+        push    .scan_x1
+        push    esi               ;[.tex_ptr]
+        push    cx
+        sar     ebx,ROUND
+        push    bx
+        sar     eax,ROUND
+        push    ax
+        call    textured_line_z
+        popad
+        mov     edx,.dz13
+        add     .cz1,edx
+        mov     edx,.dz23
+        add     .cz2,edx
+        mov     edx, .tex_dx13
+        add     .scan_x1, edx
+        mov     edx, .tex_dx23
+        add     .scan_x2, edx
+        mov     edx, .tex_dy13
+        add     .scan_y1, edx
+        mov     edx, .tex_dy23
+        add     .scan_y2, edx
+        add     eax, .dx13
+        add     ebx, .dx23
+        inc     cx
+        cmp     cx,.y3
+        jl      .tt_loop2
+.tt_loop2_end:
+.tt_end:
+        mov esp,ebp
+ret 18
+textured_line_z:
+;-----in -edi screen buffer pointer
+;------------ stack:
+  .x1 equ word [ebp+4]
+  .x2 equ word [ebp+6]
+  .y  equ word [ebp+8]
+  .tex_ptr equ dword [ebp+10]
+  .tex_x1  equ ebp+14
+  .tex_y1  equ ebp+18
+  .tex_x2  equ ebp+22
+  .tex_y2  equ ebp+26
+  .z2      equ dword [ebp+30]     ;z1, z2  coords shifted shl CATMULL_SHIFT
+  .z1      equ dword [ebp+34]
+  .z_ptr   equ dword [ebp+38]
+  .tex_dy  equ dword [ebp-4]
+  .tex_dx  equ dword [ebp-8]
+  .dz      equ dword [ebp-12]
+  .cz      equ dword [ebp-16]
+  .c_tex_x equ dword [ebp-20]  ; current tex x
+  .m_sft1 equ ebp-28
+  .m_sft2 equ ebp-32
+;  .c_tex_xM equ ebp+14
+  .tex_dxM  equ ebp-8
+        mov     ebp,esp
+        mov     ax,.y
+        or      ax,ax
+        jl      .tl_quit
+        mov     bx,[size_y_var]
+        dec     bx
+        cmp     ax,bx  ;SIZE_Y
+        jge     .tl_quit
+        mov     ax,.x1
+        cmp     ax,.x2
+        je      .tl_quit
+        jl      .tl_ok
+        xchg    ax,.x2    ; sort params
+        mov     .x1,ax
+if Ext >= MMX
+        movq    mm0,[.tex_x1]
+        movq    mm1,[.tex_x2]
+        movq    [.tex_x2],mm0
+        movq    [.tex_x1],mm1
+else
+        mov     eax,dword[.tex_x1]
+        xchg    eax,dword[.tex_x2]
+        mov     dword[.tex_x1],eax
+        mov     eax,dword[.tex_y1]
+        xchg    eax,dword[.tex_y2]
+        mov     dword[.tex_y1],eax
+end if
+        mov     eax,.z1
+        xchg    eax,.z2
+        mov     .z1,eax
+     .tl_ok:
+        mov     cx,[size_x_var]
+        dec     cx
+        cmp     .x1,cx   ;SIZE_X
+        jge     .tl_quit
+        cmp     .x2,0
+        jle     .tl_quit
+        mov     bx,.x2
+        sub     bx,.x1
+        movsx   ebx,bx
+        mov     eax,dword[.tex_y2]       ; calc .dty
+        sub     eax,dword[.tex_y1]
+        cdq
+        idiv    ebx
+        push    eax
+        mov     eax,dword[.tex_x2]       ; calc .dtx
+        sub     eax,dword[.tex_x1]
+        cdq
+        idiv    ebx
+        push    eax
+        mov     eax,.z2       ; calc .dz
+        sub     eax,.z1
+        cdq
+        idiv    ebx
+        push    eax
+        cmp    .x1,0          ; clipping
+        jg     @f
+        movsx   ebx,.x1
+        neg     ebx
+        imul    ebx           ; eax = .dz * abs(.x1)
+        add     .z1,eax
+        mov     .x1,0
+        mov     eax,.tex_dy
+        imul    ebx
+        add     dword[.tex_y1],eax
+        mov     eax,.tex_dx
+        imul    ebx
+        add     dword[.tex_x1],eax
+      @@:
+        cmp     .x2,cx ;SIZE_X
+        jl      @f
+        mov     .x2,cx  ;SIZE_X
+      @@:
+        movsx   ebx,.y        ; calc mem begin in buffers
+        movzx   eax,word[size_x_var]  ;SIZE_X
+        mul     ebx
+        movsx   ebx,.x1
+        add     eax,ebx
+        mov     ebx,eax
+        lea     eax,[eax*3]
+        add     edi,eax               ; edi - scr buff
+        shl     ebx,2
+        add     .z_ptr,ebx            ; z buffer pointer
+        mov     cx,.x2
+        sub     cx,.x1
+        movzx   ecx,cx
+;if Ext >= MMX
+;        movq    mm0,[.tex_x1]
+;        movq    mm4,mm0
+;        movq    mm1,qword[.tex_dxM]
+;        mov     ebx,.z1
+;        mov     eax,.dz
+;else
+        mov     eax,dword[.tex_x1]
+        mov     ebx,dword[.tex_y1]
+        push    .z1           ; .cz
+        push    eax           ;.c_tex_x
+;end if
+        mov     edx,.z_ptr
+     .tl_loop:
+;if Ext >= MMX
+;        cmp     ebx,[edx]                       ;  ebx - current z
+;        jge     @f
+;        movq    mm2,mm0
+;        psrad   mm2,ROUND
+;        movq    mm3,mm2
+;        psrlq   mm2,32-TEX_SHIFT
+;        paddd   mm3,mm2
+;        movd    esi,mm3
+;        mov     dword[edx],ebx    ; renew z buffer
+;else
+                                                        ;  eax - temp
+        mov     eax,.cz                         ;  ebx - cur tex y shl ROUND
+        cmp     eax,[edx]                       ;  ecx - l.lenght
+        jge     @f         ; ebx - cur tex_y ;  edx - temp
+        mov     esi,ebx                 ;  edi - scr buff
+        sar     esi,ROUND               ;  esi - tex_ptr temp
+        shl     esi,TEX_SHIFT           ;  .z_ptr - cur pointer to z buff
+        mov     eax,.c_tex_x            ;  .cz - cur z coord shl CATMULL_SHIFT
+        sar     eax,ROUND
+        add     esi,eax
+        mov     eax,.cz
+        mov     dword[edx],eax    ; renew z buffer
+;end if
+        and     esi,TEXTURE_SIZE
+        lea     esi,[esi*3]
+        add     esi,.tex_ptr
+        movsd
+        dec     edi
+        jmp     .no_skip
+      @@:
+        add     edi,3
+     .no_skip:
+        add     edx,4
+;if Ext >= MMX
+;        add     ebx,eax
+;        paddd   mm0,mm1
+;else
+        mov     eax,.dz
+        add     .cz,eax
+        mov     eax,.tex_dx
+        add     .c_tex_x,eax
+        add     ebx,.tex_dy
+;end if
+        loop    .tl_loop
+ .tl_quit:
+        mov     esp,ebp
+ret 30+8

 /programs/demos/view3ds/two_tex.inc
 ,1105 → 1,1105
-;SIZE_X equ 350
-;SIZE_Y equ 350
-;ROUND equ 8
-;TEX_X equ 512
-;TEX_Y equ 512
-;TEXTURE_SIZE EQU (512*512)-1
-;TEX_SHIFT EQU 9
-;CATMULL_SHIFT equ 8
-;TEXTURE_SIZE EQU (TEX_X * TEX_Y)-1
-;Ext = SSE
-;SSE = 3
-;MMX = 1
-;NON = 0
-;use32
-;------- Big thanks to Majuma (www.majuma.xt.pl) for absolutely great---
-;------- DOS 13h mode demos --------------------------------------------
-;------- Procedure draws triangle with two overlapped textures, I use --
-;--------Catmull Z-buffer algorithm- (Z coordinate interpolation)-------
-;--------I calc texture pixel by this way: col1*col2/256 ---------------
-two_tex_triangle_z:
-;------------------in - eax - x1 shl 16 + y1 -----------
-;---------------------- ebx - x2 shl 16 + y2 -----------
-;---------------------- ecx - x3 shl 16 + y3 -----------
-;---------------------- edx - pointer to b. texture-----
-;---------------------- esi - pointer to e. texture-----
-;---------------------- edi - pointer to screen buffer--
-;---------------------- stack : b. tex coordinates------
-;----------------------         e. tex coordinates------
-;----------------------         Z position coordinates--
-;----------------------         pointer io Z buffer-----
-;-- Z-buffer - filled with coordinates as dword --------
-;-- (Z coor. as word) shl CATMULL_SHIFT ----------------
-.b_x1   equ ebp+4   ; procedure don't save registers !!!
-.b_y1   equ ebp+6   ; each coordinate as word
-.b_x2   equ ebp+8
-.b_y2   equ ebp+10       ; b - first texture
-.b_x3   equ ebp+12
-.b_y3   equ ebp+14       ; e - second texture
-.e_x1   equ ebp+16
-.e_y1   equ ebp+18
-.e_x2   equ ebp+20
-.e_y2   equ ebp+22
-.e_x3   equ ebp+24
-.e_y3   equ ebp+26
-.z1     equ word[ebp+28]
-.z2     equ word[ebp+30]
-.z3     equ word[ebp+32]
-.z_buff equ dword[ebp+34]       ; pointer to Z-buffer
-.t_bmap equ dword[ebp-4]        ; pointer to b. texture
-.t_emap equ dword[ebp-8]        ; pointer to e. texture
-.x1     equ word[ebp-10]
-.y1     equ word[ebp-12]
-.x2     equ word[ebp-14]
-.y2     equ word[ebp-16]
-.x3     equ word[ebp-18]
-.y3     equ word[ebp-20]
-.dx12  equ dword[ebp-24]
-.dbx12 equ dword[ebp-28]
-.dby12 equ dword[ebp-32]
-.dby12q equ     [ebp-32]
-.dex12 equ dword[ebp-36]
-.dey12 equ dword[ebp-40]
-.dey12q equ     [ebp-40]
-.dz12  equ dword[ebp-44]
-.dx13  equ dword[ebp-48]
-.dbx13 equ dword[ebp-52]
-.dby13 equ dword[ebp-56]
-.dby13q equ     [ebp-56]
-.dex13 equ dword[ebp-60]
-.dey13 equ dword[ebp-64]
-.dey13q equ     [ebp-64]
-.dz13  equ dword[ebp-68]
-.dx23  equ dword[ebp-72]
-.dbx23 equ dword[ebp-76]
-.dby23 equ dword[ebp-80]
-.dby23q equ     [ebp-80]
-.dex23 equ dword[ebp-84]
-.dey23 equ dword[ebp-88]
-.dey23q equ     [ebp-88]
-.dz23  equ dword[ebp-92]
-.cx1   equ dword[ebp-96]   ; current variables
-.cx2   equ dword[ebp-100]
-.cbx1  equ dword[ebp-104]
-.cby1  equ [ebp-108]
-.cex1  equ dword[ebp-112]
-.cey1  equ [ebp-116]
-.cbx2  equ dword[ebp-120]
-.cby2  equ [ebp-124]
-.cex2  equ dword[ebp-128]
-.cey2  equ [ebp-132]
-.cz1   equ dword[ebp-136]
-.cz2   equ dword[ebp-140]
-    if Ext >= MMX
-       emms
-    else
-       cld
-    end if
-       mov     ebp,esp
-       push    edx esi       ; store bump map
-;       push    esi        ; store e. map
-     ; sub     esp,120
- .sort3:                  ; sort triangle coordinates...
-       cmp     ax,bx
-       jle     .sort1
-       xchg    eax,ebx
-       mov     edx,dword[.b_x1]
-       xchg    edx,dword[.b_x2]
-       mov     dword[.b_x1],edx
-       mov     edx,dword[.e_x1]
-       xchg    edx,dword[.e_x2]
-       mov     dword[.e_x1],edx
-       mov     dx,.z1
-       xchg    dx,.z2
-       mov     .z1,dx
- .sort1:
-       cmp      bx,cx
-       jle      .sort2
-       xchg     ebx,ecx
-       mov      edx,dword[.b_x2]
-       xchg     edx,dword[.b_x3]
-       mov      dword[.b_x2],edx
-       mov      edx,dword[.e_x2]
-       xchg     edx,dword[.e_x3]
-       mov      dword[.e_x2],edx
-       mov     dx,.z2
-       xchg    dx,.z3
-       mov     .z2,dx
-       jmp      .sort3
- .sort2:
-       push     eax ebx ecx    ; store triangle coords in variables
-;       push     ebx
-;       push     ecx
-         mov      edx,80008000h  ; eax,ebx,ecx are ANDd together into edx which means that
-         and      edx,ebx        ; if *all* of them are negative a sign flag is raised
-         and      edx,ecx
-         and      edx,eax
-         test     edx,80008000h  ; Check both X&Y at once
-         jne      .loop23_done
-    ;   mov     edx,eax         ; eax,ebx,ecx are ORd together into edx which means that
-    ;   or      edx,ebx         ; if any *one* of them is negative a sign flag is raised
-    ;   or      edx,ecx
-    ;   test    edx,80000000h   ; Check only X
-    ;   jne     .loop23_done
-    ;   cmp     .x1,SIZE_X    ; {
-    ;   jg      .loop23_done
-    ;   cmp     .x2,SIZE_X     ; This can be optimized with effort
-    ;   jg      .loop23_done
-    ;   cmp     .x3,SIZE_X
-    ;   jg      .loop23_done    ; {
-       mov      bx,.y2       ; calc delta 12
-       sub      bx,.y1
-       jnz      .bt_dx12_make
-       mov      ecx,6
-       xor      edx,edx
-     @@:
-       push     edx   ;dword 0
-       loop     @b
-       jmp      .bt_dx12_done
- .bt_dx12_make:
-       mov      ax,.x2
-       sub      ax,.x1
-       cwde
-       movsx    ebx,bx
-       shl      eax,ROUND
-       cdq
-       idiv     ebx
-;      mov      .dx12,eax
-       push      eax
-if Ext=SSE
-       sub       esp,16
-       cvtsi2ss  xmm3,ebx            ;rcps
-    ;   mov       eax,255
-       cvtsi2ss  xmm4,[i255d] ;eax
-       divss     xmm3,xmm4
-       rcpss     xmm3,xmm3
-    ;   mulss     xmm3,xmm4
-       shufps    xmm3,xmm3,0
-       movd      mm0,[.b_x1]
-       movd      mm1,[.b_x2]
-       movd      mm2,[.e_x1]
-       movd      mm3,[.e_x2]
-     ;  psubsw    mm3,mm2
-     ;  psubsw    mm1,mm0
-       pxor      mm4,mm4
-       punpcklwd mm0,mm4
-       punpcklwd mm1,mm4
-       punpcklwd mm2,mm4
-       punpcklwd mm3,mm4
-   ;    pslld     mm0,ROUND
-   ;    pslld     mm1,ROUND
-   ;    pslld     mm2,ROUND
-   ;    pslld     mm3,ROUND
-       cvtpi2ps  xmm0,mm0
-       movlhps   xmm0,xmm0
-       cvtpi2ps  xmm0,mm2
-       cvtpi2ps  xmm1,mm1
-       movlhps   xmm1,xmm1
-       cvtpi2ps  xmm1,mm3
-       subps     xmm1,xmm0
-     ;  pxor      mm4,mm4
-     ;  movq      mm5,mm1
-     ;  movq      mm6,mm1
-     ;  pcmpeqb   mm5,mm4
-;       psubd     mm1,mm0
-;       psubd     mm3,mm2
-    ;   movq      mm0,[.b_x1]      ; bx1  by1   bx2    by2
-    ;   movq      mm1,[.e_x1]      ; ex1  ey1   ex2    ey2
-    ;   pxor
-    ;   punpcklhd mm0,mm1   ; lwd  ;
-    ;   psubw     mm1,mm0   ; mm1, mm0
-    ;   pxor      mm2,mm2
-  ;     pmovmaskb eax,mm1
-  ;     and       eax,10101010b
-    ;   pcmpgtw   mm2,mm1
-    ;   punpcklwd mm1,mm2
- ;      psllw     mm0,ROUND
- ;      psllw     mm1,ROUND
- ;      movq      mm2,mm0
- ;      psrlq     mm0,32
-;       cvtpi2ps  xmm0,mm1
-;       movlhps   xmm0,xmm0
-;       cvtpi2ps  xmm0,mm3
-  ;     divps     xmm1,xmm3
-       mulps     xmm1,xmm3
-       shufps    xmm1,xmm1,10110001b
-       cvtps2pi  mm0,xmm1          ; mm0 -> 2 delta dwords
-       movhlps   xmm1,xmm1
-       cvtps2pi  mm1,xmm1
-       movq      .dey12q,mm0
-       movq      .dby12q,mm1
-;       movd      .dex12,mm0
-;       psrlq     mm0,32
-;       movd      .dey12,mm0
-;       movhlps   xmm1,xmm1
-;       cvtps2pi  mm0,xmm1
-;       movd      .dbx12,mm0
-;       psrlq     mm0,32
-;       movd      .dby12,mm0
-else
-       mov      ax,word[.b_x2]
-       sub      ax,word[.b_x1]
-       cwde
-       shl      eax,ROUND
-       cdq
-       idiv     ebx
- ;     mov      .dbx12,eax
-       push      eax
-       mov      ax,word[.b_y2]
-       sub      ax,word[.b_y1]
-       cwde
-       shl      eax,ROUND
-       cdq
-       idiv     ebx
- ;     mov      .dby12,eax
-       push      eax
- ;     mov       eax,.dbx12
- ;     mov       ebx,.dby12
- ;     int3
-       mov      ax,word[.e_x2]
-       sub      ax,word[.e_x1]
-       cwde
-       shl      eax,ROUND
-       cdq
-       idiv     ebx
- ;      mov      .dex12,eax
-       push      eax
-       mov      ax,word[.e_y2]
-       sub      ax,word[.e_y1]
-       cwde
-       shl      eax,ROUND
-       cdq
-       idiv     ebx
- ;      mov      .dey12,eax
-       push      eax
-end if
-        mov     ax,.z2
-        sub     ax,.z1
-        cwde
-        shl     eax,CATMULL_SHIFT
-        cdq
-        idiv    ebx
-        push    eax
-   .bt_dx12_done:
-       mov      bx,.y3       ; calc delta13
-       sub      bx,.y1
-       jnz      .bt_dx13_make
-       mov      ecx,6
-       xor      edx,edx
-     @@:
-       push     edx   ;dword 0
-       loop     @b
-       jmp      .bt_dx13_done
- .bt_dx13_make:
-       mov      ax,.x3
-       sub      ax,.x1
-       cwde
-       movsx    ebx,bx
-       shl      eax,ROUND
-       cdq
-       idiv     ebx
- ;      mov      .dx13,eax
-       push      eax
-if Ext=SSE
-       cvtsi2ss  xmm3,ebx
-  ;     mov       eax,255
-       cvtsi2ss  xmm4,[i255d]
-       divss     xmm3,xmm4
-       rcpss     xmm3,xmm3
-;       mulss     xmm3,xmm4
-       shufps    xmm3,xmm3,0
-       sub       esp,16
-       movd      mm0,[.b_x1]
-       movd      mm1,[.b_x3]
-       movd      mm2,[.e_x1]
-       movd      mm3,[.e_x3]
-       pxor      mm4,mm4
-       punpcklwd mm0,mm4
-       punpcklwd mm1,mm4
-       punpcklwd mm2,mm4
-       punpcklwd mm3,mm4
-       cvtpi2ps  xmm0,mm0
-       movlhps   xmm0,xmm0
-       cvtpi2ps  xmm0,mm2
-       cvtpi2ps  xmm1,mm1
-       movlhps   xmm1,xmm1
-       cvtpi2ps  xmm1,mm3
-       subps     xmm1,xmm0
-   ;    divps     xmm1,xmm3
-       mulps     xmm1,xmm3
-       shufps    xmm1,xmm1,10110001b
-       cvtps2pi  mm0,xmm1          ; mm0 -> 2 delta dwords
-       movhlps   xmm1,xmm1
-       cvtps2pi  mm1,xmm1
-       movq      .dey13q,mm0
-       movq      .dby13q,mm1
-else
-       mov      ax,word[.b_x3]
-       sub      ax,word[.b_x1]
-       cwde
-       shl      eax,ROUND
-       cdq
-       idiv     ebx
- ;      mov      .dbx13,eax
-       push      eax
-       mov      ax,word[.b_y3]
-       sub      ax,word[.b_y1]
-       cwde
-       shl      eax,ROUND
-       cdq
-       idiv     ebx
- ;      mov      .dby13,eax
-       push      eax
-       mov      ax,word[.e_x3]
-       sub      ax,word[.e_x1]
-       cwde
-       shl      eax,ROUND
-       cdq
-       idiv     ebx
- ;      mov      .dex13,eax
-       push      eax
-       mov      ax,word[.e_y3]
-       sub      ax,word[.e_y1]
-       cwde
-       shl      eax,ROUND
-       cdq
-       idiv     ebx
- ;      mov      .dey13,eax
-       push      eax
-end if
-       mov     ax,.z3
-       sub     ax,.z1
-       cwde
-       shl     eax,CATMULL_SHIFT
-       cdq
-       idiv    ebx
-  ;    mov    .dz13,eax
-       push    eax
-   .bt_dx13_done:
-       mov      bx,.y3       ; calc delta23
-       sub      bx,.y2
-       jnz      .bt_dx23_make
-       mov      ecx,6
-       xor      edx,edx
-     @@:
-       push     edx   ;dword 0
-       loop     @b
-       jmp      .bt_dx23_done
- .bt_dx23_make:
-       mov      ax,.x3
-       sub      ax,.x2
-       cwde
-       movsx    ebx,bx
-       shl      eax,ROUND
-       cdq
-       idiv     ebx
- ;      mov      .dx23,eax
-       push      eax
-if Ext=SSE
-       cvtsi2ss  xmm3,ebx
-     ;  mov       eax,255
-       cvtsi2ss  xmm4,[i255d] ;eax
-       divss     xmm3,xmm4
-       shufps    xmm3,xmm3,0
-       sub       esp,16
-       movd      mm0,[.b_x2]
-       movd      mm1,[.b_x3]
-       movd      mm2,[.e_x2]
-       movd      mm3,[.e_x3]
-       pxor      mm4,mm4
-       punpcklwd mm0,mm4
-       punpcklwd mm1,mm4
-       punpcklwd mm2,mm4
-       punpcklwd mm3,mm4
-       cvtpi2ps  xmm0,mm0
-       movlhps   xmm0,xmm0
-       cvtpi2ps  xmm0,mm2
-       cvtpi2ps  xmm1,mm1
-       movlhps   xmm1,xmm1
-       cvtpi2ps  xmm1,mm3
-       subps     xmm1,xmm0
-       divps     xmm1,xmm3
-       shufps    xmm1,xmm1,10110001b
-       cvtps2pi  mm0,xmm1          ; mm0 -> 2 delta dwords
-       movhlps   xmm1,xmm1
-       cvtps2pi  mm1,xmm1
-       movq      .dey23q,mm0
-       movq      .dby23q,mm1
-else
-       mov      ax,word[.b_x3]
-       sub      ax,word[.b_x2]
-       cwde
-       shl      eax,ROUND
-       cdq
-       idiv     ebx
- ;      mov      .dbx23,eax
-       push      eax
-       mov      ax,word[.b_y3]
-       sub      ax,word[.b_y2]
-       cwde
-       shl      eax,ROUND
-       cdq
-       idiv     ebx
- ;      mov      .dby23,eax
-       push      eax
-       mov      ax,word[.e_x3]
-       sub      ax,word[.e_x2]
-       cwde
-       shl      eax,ROUND
-       cdq
-       idiv     ebx
- ;      mov      .dex23,eax
-       push      eax
-       mov      ax,word[.e_y3]
-       sub      ax,word[.e_y2]
-       cwde
-       shl      eax,ROUND
-       cdq
-       idiv     ebx
- ;      mov      .dey23,eax
-       push      eax
-end if
-       mov     ax,.z3
-       sub     ax,.z2
-       cwde
-       shl     eax,CATMULL_SHIFT
-       cdq
-       idiv    ebx
-      ; mov     .dz23,eax
-       push    eax
-      ;  sub     esp,40
-   .bt_dx23_done:
-       movsx    eax,.x1
-       shl      eax,ROUND
-     ;  mov      .cx1,eax
-     ;  mov      .cx2,eax
-       push     eax eax
-     ; push     eax
-       movsx    eax,word[.b_x1]
-       shl      eax,ROUND
-       mov      .cbx1,eax
-       mov      .cbx2,eax
-     ;  push     eax eax
-     ; push     eax
-       movsx    eax,word[.b_y1]
-       shl      eax,ROUND
-       mov      .cby1,eax
-       mov      .cby2,eax
-     ;  push     eax eax
-     ; push     eax
-       movsx    eax,word[.e_x1]
-       shl      eax,ROUND
-       mov      .cex1,eax
-       mov      .cex2,eax
-      ; push     eax eax
-      ;push     eax
-       movsx    eax,word[.e_y1]
-       shl      eax,ROUND
-       mov      .cey1,eax
-       mov      .cey2,eax
-       sub      esp,32
-      ; push     eax eax
-      ;push     eax
-       movsx    eax,.z1
-       shl      eax,CATMULL_SHIFT
-      ; mov      .cz1,eax
-      ; mov      .cz2,eax
-      push     eax eax
-      ;push     eax
-       movsx    ecx,.y1
-       cmp      cx,.y2
-       jge      .loop12_done
-  .loop12:
-       call     .call_line
-       mov      eax,.dx13
-       add      .cx1,eax
-       mov      ebx,.dx12
-       add      .cx2,ebx
-if Ext>= SSE2
-       movups   xmm0,.cey1
-       movups   xmm1,.cey2
-       movups   xmm2,.dey12q
-       movups   xmm3,.dey13q
-       paddd    xmm0,xmm3
-       paddd    xmm1,xmm2
-       movups   .cey1,xmm0
-       movups   .cey2,xmm1
-else if (Ext = MMX) | (Ext=SSE)
-       movq     mm0,.cby2       ; with this optimization object
-       movq     mm1,.cby1       ; looks bit annoying
-       movq     mm2,.cey2
-       movq     mm3,.cey1
-       paddd    mm0,.dby12q
-       paddd    mm1,.dby13q
-       paddd    mm2,.dey12q
-       paddd    mm3,.dey13q
-       movq     .cby2,mm0
-       movq     .cby1,mm1
-       movq     .cey1,mm3
-       movq     .cey2,mm2
-else
-       mov      edx,.dbx13
-       add      .cbx1,edx
-       mov      eax,.dbx12
-       add      .cbx2,eax
-       mov      ebx,.dby13
-       add      .cby1,ebx
-       mov      edx,.dby12
-       add      .cby2,edx
-       mov      eax,.dex13
-       add      .cex1,eax
-       mov      ebx,.dex12
-       add      .cex2,ebx
-       mov      edx,.dey13
-       add      .cey1,edx
-       mov      eax,.dey12
-       add      .cey2,eax
-end if
-       mov      ebx,.dz13
-       add      .cz1,ebx
-       mov      edx,.dz12
-       add      .cz2,edx
-       inc      ecx
-       cmp      cx,.y2
-       jl       .loop12
-    .loop12_done:
-       movsx    ecx,.y2
-       cmp      cx,.y3
-       jge      .loop23_done
-       movsx    eax,.z2
-       shl      eax,CATMULL_SHIFT
-       mov      .cz2,eax
-       movsx    eax,.x2
-       shl      eax,ROUND
-       mov      .cx2,eax
-       movzx    eax,word[.b_x2]
-       shl      eax,ROUND
-       mov      .cbx2,eax
-       movzx    eax,word[.b_y2]
-       shl      eax,ROUND
-       mov      .cby2,eax
-       movzx    eax,word[.e_x2]
-       shl      eax,ROUND
-       mov      .cex2,eax
-       movzx    eax,word[.e_y2]
-       shl      eax,ROUND
-       mov      .cey2,eax
-     .loop23:
-       call     .call_line
-;if Ext = NON
-       mov      eax,.dx13
-       add      .cx1,eax
-       mov      ebx,.dx23
-       add      .cx2,ebx
-if Ext>= SSE2
-       movups   xmm0,.cey1
-       movups   xmm1,.cey2
-       movups   xmm2,.dey23q
-       movups   xmm3,.dey13q
-       paddd    xmm0,xmm3
-       paddd    xmm1,xmm2
-       movups   .cey1,xmm0
-       movups   .cey2,xmm1
-else if (Ext = MMX) | ( Ext = SSE)
-       movq     mm0,.cby2        ;    with this mmx optimization object looks bit
-       movq     mm1,.cby1        ;    annoying
-       movq     mm2,.cey2
-       movq     mm3,.cey1
-       paddd    mm0,.dby23q
-       paddd    mm1,.dby13q
-       paddd    mm2,.dey23q
-       paddd    mm3,.dey13q
-       movq     .cby2,mm0
-       movq     .cby1,mm1
-       movq     .cey2,mm2
-       movq     .cey1,mm3
-else
-       mov      edx,.dbx13
-       add      .cbx1,edx
-       mov      eax,.dbx23
-       add      .cbx2,eax
-       mov      ebx,.dby13
-       add      .cby1,ebx
-       mov      edx,.dby23
-       add      .cby2,edx
-       mov      eax,.dex13
-       add      .cex1,eax
-       mov      ebx,.dex23
-       add      .cex2,ebx
-       mov      edx,.dey13
-       add      .cey1,edx
-       mov      eax,.dey23
-       add      .cey2,eax
-end if
-       mov      ebx,.dz13
-       add      .cz1,ebx
-       mov      edx,.dz23
-       add      .cz2,edx
-;else
-;       movq     mm0,.db13q
-;       movq     mm1,.cbx1q
-       inc      ecx
-       cmp      cx,.y3
-       jl       .loop23
-    .loop23_done:
-       mov      esp,ebp
-ret   34
-.call_line:
-       pushad
-       push     .cz1
-       push     .cz2
-       push     .z_buff
-       push     .t_bmap
-       push     .t_emap
-       push     dword .cey2
-       push     .cex2
-       push     dword .cey1
-       push     .cex1
-       push     dword .cby2
-       push     .cbx2
-       push     dword .cby1
-       push     .cbx1
-       push     ecx
-       mov      eax,.cx1
-       sar      eax,ROUND
-       mov      ebx,.cx2
-       sar      ebx,ROUND
-       call     two_tex_line_z
-       popad
-ret
-two_tex_line_z:
-;--------------in: eax - x1
-;--------------    ebx - x2
-;--------------    edi - pointer to screen buffer
-;stack - another parameters :
-.y      equ dword [ebp+4]
-.bx1    equ  [ebp+8]   ;   ---
-.by1    equ  [ebp+12]  ;       |
-.bx2    equ  [ebp+16]  ;       |
-.by2    equ  [ebp+20]  ;       |>   b. texture and e. texture coords
-.ex1    equ  [ebp+24]  ;       |>   shifted shl ROUND
-.ey1    equ  [ebp+28]  ;       |
-.ex2    equ  [ebp+32]  ;       |
-.ey2    equ  [ebp+36]  ;   ---
-.emap   equ  [ebp+40]  ; b texture offset
-.bmap   equ  [ebp+44]  ; e texture offset
-.z_buff equ dword [ebp+48]
-.z2     equ dword [ebp+52]  ;   -- |>   z coords shifted
-.z1     equ dword [ebp+56]  ;   --       shl  CATMULL_SHIFT
-.x1     equ dword [ebp-4]
-.x2     equ dword [ebp-8]
-.dbx    equ [ebp-12]
-.dex    equ [ebp-16]
-.dby    equ [ebp-20]
-.dey    equ [ebp-24]
-.dz     equ dword [ebp-28]
-.cbx    equ [ebp-32]
-.cex    equ [ebp-36]
-.cby    equ [ebp-40]
-.cey    equ [ebp-44]
-.cz     equ dword [ebp-48]
-.czbuff equ dword [ebp-52]
-        mov     ebp,esp
-        mov     ecx,.y
-        or      ecx,ecx
-        jl      .bl_end
-        mov     dx,word[size_y_var]
-        dec     dx
-        cmp     cx,dx   ;word[size_y_var]  ;SIZE_Y
-        jge     .bl_end
-        cmp     eax,ebx
-        jl      @f
-        je      .bl_end
-        xchg    eax,ebx
-if Ext=NON
-        mov     edx,.bx1
-        xchg    edx,.bx2
-        mov     .bx1,edx
-        mov     edx,.by1
-        xchg    edx,.by2
-        mov     .by1,edx
-        mov     edx,.ex1
-        xchg    edx,.ex2
-        mov     .ex1,edx
-        mov     edx,.ey1
-        xchg    edx,.ey2
-        mov     .ey1,edx
-else
-        movq    mm0,.bx1
-        movq    mm1,.ex1
-        movq    mm2,.bx2
-        movq    mm3,.ex2
-        movq    .bx2,mm0
-        movq    .ex2,mm1
-        movq    .bx1,mm2
-        movq    .ex1,mm3
-end if
-        mov     edx,.z1
-        xchg    edx,.z2
-        mov     .z1,edx
-    @@:
-        push    eax ebx    ;store x1, x2
-        mov     ebx,.x1
-        movzx   edx,word[size_x_var]
-        dec     edx
-        cmp     ebx,edx
-  ;      cmp     bx,word[size_x_var]  ;SIZE_X
-        jg      .bl_end
-        cmp     .x2,0
-        jle     .bl_end
-        mov     ebx,.x2
-        sub     ebx,.x1
-if  Ext >= SSE
-       sub       esp,16
-       cvtsi2ss  xmm3,ebx            ;rcps
-       shufps    xmm3,xmm3,0
-  ;     movq      mm0,.bx1q
-  ;     movq      mm1,.bx2q
-  ;     movq      mm2,.ex1q
-  ;     movq      mm3,.ex2q
-  ;     psubd     mm1,mm0
-  ;     psubd     mm3,mm2
-  ;     cvtpi2ps  xmm1,mm1
-  ;     movlhps   xmm1,xmm1
-  ;     cvtpi2ps  xmm1,mm3
-       cvtpi2ps  xmm0,.bx1 ;mm0     ; bx1; by1
-       movlhps   xmm0,xmm0
-       cvtpi2ps  xmm0,.ex1 ;mm2     ; ex1; ey1
-       cvtpi2ps  xmm1,.bx2 ;mm1     ; bx2; by2
-       movlhps   xmm1,xmm1
-       cvtpi2ps  xmm1,.ex2 ;mm3     ; ex2; ey2
-       subps     xmm1,xmm0
-                                    ; hi             lo
-       divps     xmm1,xmm3 ; xmm1 -> dby; dbx; dey; dex
-       shufps    xmm1,xmm1,11011000b
-       cvtps2pi  mm0,xmm1          ; mm0 -> 2 delta dwords
-       movhlps   xmm1,xmm1
-       cvtps2pi  mm1,xmm1
-       movq      .dex,mm0 ; hi - lo  ->  dbx, dex
-       movq      .dey,mm1 ; hi - lo  ->  dby, dey
-else
-        mov     eax,.bx2       ; calc .dbx
-        sub     eax,.bx1
-        cdq
-        idiv    ebx
-        push    eax
-        mov     eax,.ex2       ; calc .dby
-        sub     eax,.ex1
-        cdq
-        idiv    ebx
-        push    eax
-        mov     eax,.by2       ; calc .dex
-        sub     eax,.by1
-        cdq
-        idiv    ebx
-        push    eax
-        mov     eax,.ey2       ; calc .dey
-        sub     eax,.ey1
-        cdq
-        idiv    ebx
-        push    eax
-end if
-        mov     eax,.z2        ; calc .dz
-        sub     eax,.z1
-        cdq
-        idiv    ebx
-        push    eax
-        cmp     .x1,0         ; set correctly begin variable
-        jge     @f            ; CLIPPING ON FUNCTION
-                              ; cutting triangle exceedes screen
-        mov     ebx,.x1
-        neg     ebx
-        imul    ebx           ; eax = .dz * abs(.x1)
-        add     .z1,eax
-        mov     .x1,0
-        mov     eax,.dbx
-        imul    ebx
-        add    .bx1,eax
-        mov     eax,.dby
-        imul    ebx
-        add     .by1,eax
-        mov     eax,.dex
-        imul    ebx
-        add     .ex1,eax
-        mov     eax,.dey
-        imul    ebx
-        add     .ey1,eax
-      @@:
-        movzx   eax,word[size_x_var] ;SIZE_X  ;word[size_x_var]
-        mov     ebx,.x2
-        cmp     eax,ebx
-        jg      @f
-        mov     .x2,eax
-      @@:
-      ;  movd    mm0,eax
-      ;  movd    mm1,.x2
-      ;  pminsw  mm0,mm1
-      ;  movd    .x2,mm0
- ;       cmp     .x2,SIZE_X  ;eax   |
- ;       jl      @f                 |>       this dont work idk cause
- ;       mov     .x2,SIZE_X ;eax    |
-      @@:
-      ;  movzx   eax,word[size_x_var]       ;calc memory begin in buffers
-        mov     ebx,.y
-        mul     ebx
-        mov     ebx,.x1
-        add     eax,ebx
-        mov     ebx,eax
-        lea     eax,[eax*3]
-        add     edi,eax           ; edi - screen
-        mov     esi,.z_buff       ; z-buffer filled with dd variables
-        shl     ebx,2
-        add     esi,ebx           ; esi - Z buffer
-        mov     ecx,.x2
-        sub     ecx,.x1
-        ; init current variables
-        push    dword .bx1 ;.by1 .ex1 .ey1 .z1 esi
-        push    dword .ex1
-        push    dword .by1
-        push    dword .ey1
-        push    .z1              ; current z shl CATMULL_SHIFT
-        push    esi
-if Ext >= MMX
-     pxor   mm0,mm0
-     movq   mm3,.cex   ; hi - lo -> cbx; cex
-     movq   mm4,.cey   ; hi - lo -> cby; cey
-;     movq   mm5,mm3
-;     movq   mm6,mm4
-;     psrad  mm5,ROUND
-;     psrad  mm6,ROUND
-;     movq   .ceyq,mm5
-;     movq   .cbyq,mm6
-     mov    edx,.czbuff
-else
-     cld
-end if
-     .draw:
-    ; if TEX = SHIFTING   ;bump drawing only in shifting mode
-if Ext=NON
-        mov     esi,.czbuff      ; .czbuff current address in buffer
-        mov     ebx,.cz          ; .cz - cur z position
-        cmp     ebx,dword[esi]
-else
-        mov     ebx,.cz
-        cmp     ebx,dword[edx]
-end if
-        jge     .skip
-if Ext=NON
-        mov     eax,.cby
-        mov     esi,.cbx
-        sar     eax,ROUND
-        sar     esi,ROUND
-        shl     eax,TEX_SHIFT   ;-
-        add     esi,eax
-        lea     esi,[esi*3]         ;-  ; esi - current b. texture addres
-        add     esi,.bmap
-        mov     ebx,.cex       ;.cex - current env map X
-        mov     eax,.cey       ;.cey - current  env map y
-        sar     ebx,ROUND
-        sar     eax,ROUND
-        shl     eax,TEX_SHIFT
-        add     ebx,eax
-        lea     ebx,[ebx*3]
-        add     ebx,.emap
-else
-        movq    mm5,mm4 ;.cey
-        psrad   mm5,ROUND
-        pslld   mm5,TEX_SHIFT
-        movq    mm6,mm3 ;.cex
-        psrad   mm6,ROUND
-        paddd   mm5,mm6
-        movq    mm6,mm5
-        paddd   mm5,mm5
-        paddd   mm5,mm6
-        paddd   mm5,.emap
-        movd    esi,mm5
-        psrlq   mm5,32
-        movd    ebx,mm5
-end if
-if Ext>=MMX
-        movd      mm1,[esi]
-        movd      mm2,[ebx]
-        punpcklbw mm1,mm0
-        punpcklbw mm2,mm0
-        pmullw    mm1,mm2
-        psrlw     mm1,8
-        packuswb  mm1,mm0
-        movd      [edi],mm1
-        mov       ebx,.cz
-        mov       dword[edx],ebx
-else
-        cld                     ; esi - tex e.
-        lodsb                   ; ebx - tex b.
-        mov     dl,[ebx]
-        mul     dl
-        shr     ax,8
-        stosb
-        inc     ebx
-        lodsb
-        mov     dl,[ebx]
-        mul     dl
-        shr     ax,8
-        stosb
-        inc     ebx
-        lodsb
-        mov     dl,[ebx]
-        mul     dl
-        shr     ax,8
-        stosb
-        mov     ebx,.cz
-        mov     esi,.czbuff
-        mov     dword[esi],ebx
-        jmp     .no_skip
-end if
-     .skip:
-        add     edi,3
-   if Ext = NON
-     .no_skip:
-        add     .czbuff,4
-        mov     eax,.dbx
-        add     .cbx,eax
-        mov     eax,.dby
-        add     .cby,eax
-        mov     eax,.dex
-        add     .cex,eax
-        mov     eax,.dey
-        add     .cey,eax
-    else
-        add     edx,4
-        paddd   mm3,.dex
-        paddd   mm4,.dey
-  ;      movq    mm5,mm3
-  ;      movq    mm6,mm4
-  ;      psrad   mm5,ROUND
-  ;      psrad   mm6,ROUND
-     ;   movq    .cex,mm3
-     ;   movq    .cey,mm4
-    end if
-        mov     eax,.dz
-        add     .cz,eax
-    if Ext = NON
-        dec     ecx
-        jnz     .draw
-    else
-        loop    .draw
-    end if
-  .bl_end:
-        mov     esp,ebp
-ret 56
+;SIZE_X equ 350
+;SIZE_Y equ 350
+;ROUND equ 8
+;TEX_X equ 512
+;TEX_Y equ 512
+;TEXTURE_SIZE EQU (512*512)-1
+;TEX_SHIFT EQU 9
+;CATMULL_SHIFT equ 8
+;TEXTURE_SIZE EQU (TEX_X * TEX_Y)-1
+;Ext = SSE
+;SSE = 3
+;MMX = 1
+;NON = 0
+;use32
+;------- Big thanks to Majuma (www.majuma.xt.pl) for absolutely great---
+;------- DOS 13h mode demos --------------------------------------------
+;------- Procedure draws triangle with two overlapped textures, I use --
+;--------Catmull Z-buffer algorithm- (Z coordinate interpolation)-------
+;--------I calc texture pixel by this way: col1*col2/256 ---------------
+two_tex_triangle_z:
+;------------------in - eax - x1 shl 16 + y1 -----------
+;---------------------- ebx - x2 shl 16 + y2 -----------
+;---------------------- ecx - x3 shl 16 + y3 -----------
+;---------------------- edx - pointer to b. texture-----
+;---------------------- esi - pointer to e. texture-----
+;---------------------- edi - pointer to screen buffer--
+;---------------------- stack : b. tex coordinates------
+;----------------------         e. tex coordinates------
+;----------------------         Z position coordinates--
+;----------------------         pointer io Z buffer-----
+;-- Z-buffer - filled with coordinates as dword --------
+;-- (Z coor. as word) shl CATMULL_SHIFT ----------------
+.b_x1   equ ebp+4   ; procedure don't save registers !!!
+.b_y1   equ ebp+6   ; each coordinate as word
+.b_x2   equ ebp+8
+.b_y2   equ ebp+10       ; b - first texture
+.b_x3   equ ebp+12
+.b_y3   equ ebp+14       ; e - second texture
+.e_x1   equ ebp+16
+.e_y1   equ ebp+18
+.e_x2   equ ebp+20
+.e_y2   equ ebp+22
+.e_x3   equ ebp+24
+.e_y3   equ ebp+26
+.z1     equ word[ebp+28]
+.z2     equ word[ebp+30]
+.z3     equ word[ebp+32]
+.z_buff equ dword[ebp+34]       ; pointer to Z-buffer
+.t_bmap equ dword[ebp-4]        ; pointer to b. texture
+.t_emap equ dword[ebp-8]        ; pointer to e. texture
+.x1     equ word[ebp-10]
+.y1     equ word[ebp-12]
+.x2     equ word[ebp-14]
+.y2     equ word[ebp-16]
+.x3     equ word[ebp-18]
+.y3     equ word[ebp-20]
+.dx12  equ dword[ebp-24]
+.dbx12 equ dword[ebp-28]
+.dby12 equ dword[ebp-32]
+.dby12q equ     [ebp-32]
+.dex12 equ dword[ebp-36]
+.dey12 equ dword[ebp-40]
+.dey12q equ     [ebp-40]
+.dz12  equ dword[ebp-44]
+.dx13  equ dword[ebp-48]
+.dbx13 equ dword[ebp-52]
+.dby13 equ dword[ebp-56]
+.dby13q equ     [ebp-56]
+.dex13 equ dword[ebp-60]
+.dey13 equ dword[ebp-64]
+.dey13q equ     [ebp-64]
+.dz13  equ dword[ebp-68]
+.dx23  equ dword[ebp-72]
+.dbx23 equ dword[ebp-76]
+.dby23 equ dword[ebp-80]
+.dby23q equ     [ebp-80]
+.dex23 equ dword[ebp-84]
+.dey23 equ dword[ebp-88]
+.dey23q equ     [ebp-88]
+.dz23  equ dword[ebp-92]
+.cx1   equ dword[ebp-96]   ; current variables
+.cx2   equ dword[ebp-100]
+.cbx1  equ dword[ebp-104]
+.cby1  equ [ebp-108]
+.cex1  equ dword[ebp-112]
+.cey1  equ [ebp-116]
+.cbx2  equ dword[ebp-120]
+.cby2  equ [ebp-124]
+.cex2  equ dword[ebp-128]
+.cey2  equ [ebp-132]
+.cz1   equ dword[ebp-136]
+.cz2   equ dword[ebp-140]
+    if Ext >= MMX
+       emms
+    else
+       cld
+    end if
+       mov     ebp,esp
+       push    edx esi       ; store bump map
+;       push    esi        ; store e. map
+     ; sub     esp,120
+ .sort3:                  ; sort triangle coordinates...
+       cmp     ax,bx
+       jle     .sort1
+       xchg    eax,ebx
+       mov     edx,dword[.b_x1]
+       xchg    edx,dword[.b_x2]
+       mov     dword[.b_x1],edx
+       mov     edx,dword[.e_x1]
+       xchg    edx,dword[.e_x2]
+       mov     dword[.e_x1],edx
+       mov     dx,.z1
+       xchg    dx,.z2
+       mov     .z1,dx
+ .sort1:
+       cmp      bx,cx
+       jle      .sort2
+       xchg     ebx,ecx
+       mov      edx,dword[.b_x2]
+       xchg     edx,dword[.b_x3]
+       mov      dword[.b_x2],edx
+       mov      edx,dword[.e_x2]
+       xchg     edx,dword[.e_x3]
+       mov      dword[.e_x2],edx
+       mov     dx,.z2
+       xchg    dx,.z3
+       mov     .z2,dx
+       jmp      .sort3
+ .sort2:
+       push     eax ebx ecx    ; store triangle coords in variables
+;       push     ebx
+;       push     ecx
+         mov      edx,80008000h  ; eax,ebx,ecx are ANDd together into edx which means that
+         and      edx,ebx        ; if *all* of them are negative a sign flag is raised
+         and      edx,ecx
+         and      edx,eax
+         test     edx,80008000h  ; Check both X&Y at once
+         jne      .loop23_done
+    ;   mov     edx,eax         ; eax,ebx,ecx are ORd together into edx which means that
+    ;   or      edx,ebx         ; if any *one* of them is negative a sign flag is raised
+    ;   or      edx,ecx
+    ;   test    edx,80000000h   ; Check only X
+    ;   jne     .loop23_done
+    ;   cmp     .x1,SIZE_X    ; {
+    ;   jg      .loop23_done
+    ;   cmp     .x2,SIZE_X     ; This can be optimized with effort
+    ;   jg      .loop23_done
+    ;   cmp     .x3,SIZE_X
+    ;   jg      .loop23_done    ; {
+       mov      bx,.y2       ; calc delta 12
+       sub      bx,.y1
+       jnz      .bt_dx12_make
+       mov      ecx,6
+       xor      edx,edx
+     @@:
+       push     edx   ;dword 0
+       loop     @b
+       jmp      .bt_dx12_done
+ .bt_dx12_make:
+       mov      ax,.x2
+       sub      ax,.x1
+       cwde
+       movsx    ebx,bx
+       shl      eax,ROUND
+       cdq
+       idiv     ebx
+;      mov      .dx12,eax
+       push      eax
+if Ext=SSE
+       sub       esp,16
+       cvtsi2ss  xmm3,ebx            ;rcps
+    ;   mov       eax,255
+       cvtsi2ss  xmm4,[i255d] ;eax
+       divss     xmm3,xmm4
+       rcpss     xmm3,xmm3
+    ;   mulss     xmm3,xmm4
+       shufps    xmm3,xmm3,0
+       movd      mm0,[.b_x1]
+       movd      mm1,[.b_x2]
+       movd      mm2,[.e_x1]
+       movd      mm3,[.e_x2]
+     ;  psubsw    mm3,mm2
+     ;  psubsw    mm1,mm0
+       pxor      mm4,mm4
+       punpcklwd mm0,mm4
+       punpcklwd mm1,mm4
+       punpcklwd mm2,mm4
+       punpcklwd mm3,mm4
+   ;    pslld     mm0,ROUND
+   ;    pslld     mm1,ROUND
+   ;    pslld     mm2,ROUND
+   ;    pslld     mm3,ROUND
+       cvtpi2ps  xmm0,mm0
+       movlhps   xmm0,xmm0
+       cvtpi2ps  xmm0,mm2
+       cvtpi2ps  xmm1,mm1
+       movlhps   xmm1,xmm1
+       cvtpi2ps  xmm1,mm3
+       subps     xmm1,xmm0
+     ;  pxor      mm4,mm4
+     ;  movq      mm5,mm1
+     ;  movq      mm6,mm1
+     ;  pcmpeqb   mm5,mm4
+;       psubd     mm1,mm0
+;       psubd     mm3,mm2
+    ;   movq      mm0,[.b_x1]      ; bx1  by1   bx2    by2
+    ;   movq      mm1,[.e_x1]      ; ex1  ey1   ex2    ey2
+    ;   pxor
+    ;   punpcklhd mm0,mm1   ; lwd  ;
+    ;   psubw     mm1,mm0   ; mm1, mm0
+    ;   pxor      mm2,mm2
+  ;     pmovmaskb eax,mm1
+  ;     and       eax,10101010b
+    ;   pcmpgtw   mm2,mm1
+    ;   punpcklwd mm1,mm2
+ ;      psllw     mm0,ROUND
+ ;      psllw     mm1,ROUND
+ ;      movq      mm2,mm0
+ ;      psrlq     mm0,32
+;       cvtpi2ps  xmm0,mm1
+;       movlhps   xmm0,xmm0
+;       cvtpi2ps  xmm0,mm3
+  ;     divps     xmm1,xmm3
+       mulps     xmm1,xmm3
+       shufps    xmm1,xmm1,10110001b
+       cvtps2pi  mm0,xmm1          ; mm0 -> 2 delta dwords
+       movhlps   xmm1,xmm1
+       cvtps2pi  mm1,xmm1
+       movq      .dey12q,mm0
+       movq      .dby12q,mm1
+;       movd      .dex12,mm0
+;       psrlq     mm0,32
+;       movd      .dey12,mm0
+;       movhlps   xmm1,xmm1
+;       cvtps2pi  mm0,xmm1
+;       movd      .dbx12,mm0
+;       psrlq     mm0,32
+;       movd      .dby12,mm0
+else
+       mov      ax,word[.b_x2]
+       sub      ax,word[.b_x1]
+       cwde
+       shl      eax,ROUND
+       cdq
+       idiv     ebx
+ ;     mov      .dbx12,eax
+       push      eax
+       mov      ax,word[.b_y2]
+       sub      ax,word[.b_y1]
+       cwde
+       shl      eax,ROUND
+       cdq
+       idiv     ebx
+ ;     mov      .dby12,eax
+       push      eax
+ ;     mov       eax,.dbx12
+ ;     mov       ebx,.dby12
+ ;     int3
+       mov      ax,word[.e_x2]
+       sub      ax,word[.e_x1]
+       cwde
+       shl      eax,ROUND
+       cdq
+       idiv     ebx
+ ;      mov      .dex12,eax
+       push      eax
+       mov      ax,word[.e_y2]
+       sub      ax,word[.e_y1]
+       cwde
+       shl      eax,ROUND
+       cdq
+       idiv     ebx
+ ;      mov      .dey12,eax
+       push      eax
+end if
+        mov     ax,.z2
+        sub     ax,.z1
+        cwde
+        shl     eax,CATMULL_SHIFT
+        cdq
+        idiv    ebx
+        push    eax
+   .bt_dx12_done:
+       mov      bx,.y3       ; calc delta13
+       sub      bx,.y1
+       jnz      .bt_dx13_make
+       mov      ecx,6
+       xor      edx,edx
+     @@:
+       push     edx   ;dword 0
+       loop     @b
+       jmp      .bt_dx13_done
+ .bt_dx13_make:
+       mov      ax,.x3
+       sub      ax,.x1
+       cwde
+       movsx    ebx,bx
+       shl      eax,ROUND
+       cdq
+       idiv     ebx
+ ;      mov      .dx13,eax
+       push      eax
+if Ext=SSE
+       cvtsi2ss  xmm3,ebx
+  ;     mov       eax,255
+       cvtsi2ss  xmm4,[i255d]
+       divss     xmm3,xmm4
+       rcpss     xmm3,xmm3
+;       mulss     xmm3,xmm4
+       shufps    xmm3,xmm3,0
+       sub       esp,16
+       movd      mm0,[.b_x1]
+       movd      mm1,[.b_x3]
+       movd      mm2,[.e_x1]
+       movd      mm3,[.e_x3]
+       pxor      mm4,mm4
+       punpcklwd mm0,mm4
+       punpcklwd mm1,mm4
+       punpcklwd mm2,mm4
+       punpcklwd mm3,mm4
+       cvtpi2ps  xmm0,mm0
+       movlhps   xmm0,xmm0
+       cvtpi2ps  xmm0,mm2
+       cvtpi2ps  xmm1,mm1
+       movlhps   xmm1,xmm1
+       cvtpi2ps  xmm1,mm3
+       subps     xmm1,xmm0
+   ;    divps     xmm1,xmm3
+       mulps     xmm1,xmm3
+       shufps    xmm1,xmm1,10110001b
+       cvtps2pi  mm0,xmm1          ; mm0 -> 2 delta dwords
+       movhlps   xmm1,xmm1
+       cvtps2pi  mm1,xmm1
+       movq      .dey13q,mm0
+       movq      .dby13q,mm1
+else
+       mov      ax,word[.b_x3]
+       sub      ax,word[.b_x1]
+       cwde
+       shl      eax,ROUND
+       cdq
+       idiv     ebx
+ ;      mov      .dbx13,eax
+       push      eax
+       mov      ax,word[.b_y3]
+       sub      ax,word[.b_y1]
+       cwde
+       shl      eax,ROUND
+       cdq
+       idiv     ebx
+ ;      mov      .dby13,eax
+       push      eax
+       mov      ax,word[.e_x3]
+       sub      ax,word[.e_x1]
+       cwde
+       shl      eax,ROUND
+       cdq
+       idiv     ebx
+ ;      mov      .dex13,eax
+       push      eax
+       mov      ax,word[.e_y3]
+       sub      ax,word[.e_y1]
+       cwde
+       shl      eax,ROUND
+       cdq
+       idiv     ebx
+ ;      mov      .dey13,eax
+       push      eax
+end if
+       mov     ax,.z3
+       sub     ax,.z1
+       cwde
+       shl     eax,CATMULL_SHIFT
+       cdq
+       idiv    ebx
+  ;    mov    .dz13,eax
+       push    eax
+   .bt_dx13_done:
+       mov      bx,.y3       ; calc delta23
+       sub      bx,.y2
+       jnz      .bt_dx23_make
+       mov      ecx,6
+       xor      edx,edx
+     @@:
+       push     edx   ;dword 0
+       loop     @b
+       jmp      .bt_dx23_done
+ .bt_dx23_make:
+       mov      ax,.x3
+       sub      ax,.x2
+       cwde
+       movsx    ebx,bx
+       shl      eax,ROUND
+       cdq
+       idiv     ebx
+ ;      mov      .dx23,eax
+       push      eax
+if Ext=SSE
+       cvtsi2ss  xmm3,ebx
+     ;  mov       eax,255
+       cvtsi2ss  xmm4,[i255d] ;eax
+       divss     xmm3,xmm4
+       shufps    xmm3,xmm3,0
+       sub       esp,16
+       movd      mm0,[.b_x2]
+       movd      mm1,[.b_x3]
+       movd      mm2,[.e_x2]
+       movd      mm3,[.e_x3]
+       pxor      mm4,mm4
+       punpcklwd mm0,mm4
+       punpcklwd mm1,mm4
+       punpcklwd mm2,mm4
+       punpcklwd mm3,mm4
+       cvtpi2ps  xmm0,mm0
+       movlhps   xmm0,xmm0
+       cvtpi2ps  xmm0,mm2
+       cvtpi2ps  xmm1,mm1
+       movlhps   xmm1,xmm1
+       cvtpi2ps  xmm1,mm3
+       subps     xmm1,xmm0
+       divps     xmm1,xmm3
+       shufps    xmm1,xmm1,10110001b
+       cvtps2pi  mm0,xmm1          ; mm0 -> 2 delta dwords
+       movhlps   xmm1,xmm1
+       cvtps2pi  mm1,xmm1
+       movq      .dey23q,mm0
+       movq      .dby23q,mm1
+else
+       mov      ax,word[.b_x3]
+       sub      ax,word[.b_x2]
+       cwde
+       shl      eax,ROUND
+       cdq
+       idiv     ebx
+ ;      mov      .dbx23,eax
+       push      eax
+       mov      ax,word[.b_y3]
+       sub      ax,word[.b_y2]
+       cwde
+       shl      eax,ROUND
+       cdq
+       idiv     ebx
+ ;      mov      .dby23,eax
+       push      eax
+       mov      ax,word[.e_x3]
+       sub      ax,word[.e_x2]
+       cwde
+       shl      eax,ROUND
+       cdq
+       idiv     ebx
+ ;      mov      .dex23,eax
+       push      eax
+       mov      ax,word[.e_y3]
+       sub      ax,word[.e_y2]
+       cwde
+       shl      eax,ROUND
+       cdq
+       idiv     ebx
+ ;      mov      .dey23,eax
+       push      eax
+end if
+       mov     ax,.z3
+       sub     ax,.z2
+       cwde
+       shl     eax,CATMULL_SHIFT
+       cdq
+       idiv    ebx
+      ; mov     .dz23,eax
+       push    eax
+      ;  sub     esp,40
+   .bt_dx23_done:
+       movsx    eax,.x1
+       shl      eax,ROUND
+     ;  mov      .cx1,eax
+     ;  mov      .cx2,eax
+       push     eax eax
+     ; push     eax
+       movsx    eax,word[.b_x1]
+       shl      eax,ROUND
+       mov      .cbx1,eax
+       mov      .cbx2,eax
+     ;  push     eax eax
+     ; push     eax
+       movsx    eax,word[.b_y1]
+       shl      eax,ROUND
+       mov      .cby1,eax
+       mov      .cby2,eax
+     ;  push     eax eax
+     ; push     eax
+       movsx    eax,word[.e_x1]
+       shl      eax,ROUND
+       mov      .cex1,eax
+       mov      .cex2,eax
+      ; push     eax eax
+      ;push     eax
+       movsx    eax,word[.e_y1]
+       shl      eax,ROUND
+       mov      .cey1,eax
+       mov      .cey2,eax
+       sub      esp,32
+      ; push     eax eax
+      ;push     eax
+       movsx    eax,.z1
+       shl      eax,CATMULL_SHIFT
+      ; mov      .cz1,eax
+      ; mov      .cz2,eax
+      push     eax eax
+      ;push     eax
+       movsx    ecx,.y1
+       cmp      cx,.y2
+       jge      .loop12_done
+  .loop12:
+       call     .call_line
+       mov      eax,.dx13
+       add      .cx1,eax
+       mov      ebx,.dx12
+       add      .cx2,ebx
+if Ext>= SSE2
+       movups   xmm0,.cey1
+       movups   xmm1,.cey2
+       movups   xmm2,.dey12q
+       movups   xmm3,.dey13q
+       paddd    xmm0,xmm3
+       paddd    xmm1,xmm2
+       movups   .cey1,xmm0
+       movups   .cey2,xmm1
+else if (Ext = MMX) | (Ext=SSE)
+       movq     mm0,.cby2       ; with this optimization object
+       movq     mm1,.cby1       ; looks bit annoying
+       movq     mm2,.cey2
+       movq     mm3,.cey1
+       paddd    mm0,.dby12q
+       paddd    mm1,.dby13q
+       paddd    mm2,.dey12q
+       paddd    mm3,.dey13q
+       movq     .cby2,mm0
+       movq     .cby1,mm1
+       movq     .cey1,mm3
+       movq     .cey2,mm2
+else
+       mov      edx,.dbx13
+       add      .cbx1,edx
+       mov      eax,.dbx12
+       add      .cbx2,eax
+       mov      ebx,.dby13
+       add      .cby1,ebx
+       mov      edx,.dby12
+       add      .cby2,edx
+       mov      eax,.dex13
+       add      .cex1,eax
+       mov      ebx,.dex12
+       add      .cex2,ebx
+       mov      edx,.dey13
+       add      .cey1,edx
+       mov      eax,.dey12
+       add      .cey2,eax
+end if
+       mov      ebx,.dz13
+       add      .cz1,ebx
+       mov      edx,.dz12
+       add      .cz2,edx
+       inc      ecx
+       cmp      cx,.y2
+       jl       .loop12
+    .loop12_done:
+       movsx    ecx,.y2
+       cmp      cx,.y3
+       jge      .loop23_done
+       movsx    eax,.z2
+       shl      eax,CATMULL_SHIFT
+       mov      .cz2,eax
+       movsx    eax,.x2
+       shl      eax,ROUND
+       mov      .cx2,eax
+       movzx    eax,word[.b_x2]
+       shl      eax,ROUND
+       mov      .cbx2,eax
+       movzx    eax,word[.b_y2]
+       shl      eax,ROUND
+       mov      .cby2,eax
+       movzx    eax,word[.e_x2]
+       shl      eax,ROUND
+       mov      .cex2,eax
+       movzx    eax,word[.e_y2]
+       shl      eax,ROUND
+       mov      .cey2,eax
+     .loop23:
+       call     .call_line
+;if Ext = NON
+       mov      eax,.dx13
+       add      .cx1,eax
+       mov      ebx,.dx23
+       add      .cx2,ebx
+if Ext>= SSE2
+       movups   xmm0,.cey1
+       movups   xmm1,.cey2
+       movups   xmm2,.dey23q
+       movups   xmm3,.dey13q
+       paddd    xmm0,xmm3
+       paddd    xmm1,xmm2
+       movups   .cey1,xmm0
+       movups   .cey2,xmm1
+else if (Ext = MMX) | ( Ext = SSE)
+       movq     mm0,.cby2        ;    with this mmx optimization object looks bit
+       movq     mm1,.cby1        ;    annoying
+       movq     mm2,.cey2
+       movq     mm3,.cey1
+       paddd    mm0,.dby23q
+       paddd    mm1,.dby13q
+       paddd    mm2,.dey23q
+       paddd    mm3,.dey13q
+       movq     .cby2,mm0
+       movq     .cby1,mm1
+       movq     .cey2,mm2
+       movq     .cey1,mm3
+else
+       mov      edx,.dbx13
+       add      .cbx1,edx
+       mov      eax,.dbx23
+       add      .cbx2,eax
+       mov      ebx,.dby13
+       add      .cby1,ebx
+       mov      edx,.dby23
+       add      .cby2,edx
+       mov      eax,.dex13
+       add      .cex1,eax
+       mov      ebx,.dex23
+       add      .cex2,ebx
+       mov      edx,.dey13
+       add      .cey1,edx
+       mov      eax,.dey23
+       add      .cey2,eax
+end if
+       mov      ebx,.dz13
+       add      .cz1,ebx
+       mov      edx,.dz23
+       add      .cz2,edx
+;else
+;       movq     mm0,.db13q
+;       movq     mm1,.cbx1q
+       inc      ecx
+       cmp      cx,.y3
+       jl       .loop23
+    .loop23_done:
+       mov      esp,ebp
+ret   34
+.call_line:
+       pushad
+       push     .cz1
+       push     .cz2
+       push     .z_buff
+       push     .t_bmap
+       push     .t_emap
+       push     dword .cey2
+       push     .cex2
+       push     dword .cey1
+       push     .cex1
+       push     dword .cby2
+       push     .cbx2
+       push     dword .cby1
+       push     .cbx1
+       push     ecx
+       mov      eax,.cx1
+       sar      eax,ROUND
+       mov      ebx,.cx2
+       sar      ebx,ROUND
+       call     two_tex_line_z
+       popad
+ret
+two_tex_line_z:
+;--------------in: eax - x1
+;--------------    ebx - x2
+;--------------    edi - pointer to screen buffer
+;stack - another parameters :
+.y      equ dword [ebp+4]
+.bx1    equ  [ebp+8]   ;   ---
+.by1    equ  [ebp+12]  ;       |
+.bx2    equ  [ebp+16]  ;       |
+.by2    equ  [ebp+20]  ;       |>   b. texture and e. texture coords
+.ex1    equ  [ebp+24]  ;       |>   shifted shl ROUND
+.ey1    equ  [ebp+28]  ;       |
+.ex2    equ  [ebp+32]  ;       |
+.ey2    equ  [ebp+36]  ;   ---
+.emap   equ  [ebp+40]  ; b texture offset
+.bmap   equ  [ebp+44]  ; e texture offset
+.z_buff equ dword [ebp+48]
+.z2     equ dword [ebp+52]  ;   -- |>   z coords shifted
+.z1     equ dword [ebp+56]  ;   --       shl  CATMULL_SHIFT
+.x1     equ dword [ebp-4]
+.x2     equ dword [ebp-8]
+.dbx    equ [ebp-12]
+.dex    equ [ebp-16]
+.dby    equ [ebp-20]
+.dey    equ [ebp-24]
+.dz     equ dword [ebp-28]
+.cbx    equ [ebp-32]
+.cex    equ [ebp-36]
+.cby    equ [ebp-40]
+.cey    equ [ebp-44]
+.cz     equ dword [ebp-48]
+.czbuff equ dword [ebp-52]
+        mov     ebp,esp
+        mov     ecx,.y
+        or      ecx,ecx
+        jl      .bl_end
+        mov     dx,word[size_y_var]
+        dec     dx
+        cmp     cx,dx   ;word[size_y_var]  ;SIZE_Y
+        jge     .bl_end
+        cmp     eax,ebx
+        jl      @f
+        je      .bl_end
+        xchg    eax,ebx
+if Ext=NON
+        mov     edx,.bx1
+        xchg    edx,.bx2
+        mov     .bx1,edx
+        mov     edx,.by1
+        xchg    edx,.by2
+        mov     .by1,edx
+        mov     edx,.ex1
+        xchg    edx,.ex2
+        mov     .ex1,edx
+        mov     edx,.ey1
+        xchg    edx,.ey2
+        mov     .ey1,edx
+else
+        movq    mm0,.bx1
+        movq    mm1,.ex1
+        movq    mm2,.bx2
+        movq    mm3,.ex2
+        movq    .bx2,mm0
+        movq    .ex2,mm1
+        movq    .bx1,mm2
+        movq    .ex1,mm3
+end if
+        mov     edx,.z1
+        xchg    edx,.z2
+        mov     .z1,edx
+    @@:
+        push    eax ebx    ;store x1, x2
+        mov     ebx,.x1
+        movzx   edx,word[size_x_var]
+        dec     edx
+        cmp     ebx,edx
+  ;      cmp     bx,word[size_x_var]  ;SIZE_X
+        jg      .bl_end
+        cmp     .x2,0
+        jle     .bl_end
+        mov     ebx,.x2
+        sub     ebx,.x1
+if  Ext >= SSE
+       sub       esp,16
+       cvtsi2ss  xmm3,ebx            ;rcps
+       shufps    xmm3,xmm3,0
+  ;     movq      mm0,.bx1q
+  ;     movq      mm1,.bx2q
+  ;     movq      mm2,.ex1q
+  ;     movq      mm3,.ex2q
+  ;     psubd     mm1,mm0
+  ;     psubd     mm3,mm2
+  ;     cvtpi2ps  xmm1,mm1
+  ;     movlhps   xmm1,xmm1
+  ;     cvtpi2ps  xmm1,mm3
+       cvtpi2ps  xmm0,.bx1 ;mm0     ; bx1; by1
+       movlhps   xmm0,xmm0
+       cvtpi2ps  xmm0,.ex1 ;mm2     ; ex1; ey1
+       cvtpi2ps  xmm1,.bx2 ;mm1     ; bx2; by2
+       movlhps   xmm1,xmm1
+       cvtpi2ps  xmm1,.ex2 ;mm3     ; ex2; ey2
+       subps     xmm1,xmm0
+                                    ; hi             lo
+       divps     xmm1,xmm3 ; xmm1 -> dby; dbx; dey; dex
+       shufps    xmm1,xmm1,11011000b
+       cvtps2pi  mm0,xmm1          ; mm0 -> 2 delta dwords
+       movhlps   xmm1,xmm1
+       cvtps2pi  mm1,xmm1
+       movq      .dex,mm0 ; hi - lo  ->  dbx, dex
+       movq      .dey,mm1 ; hi - lo  ->  dby, dey
+else
+        mov     eax,.bx2       ; calc .dbx
+        sub     eax,.bx1
+        cdq
+        idiv    ebx
+        push    eax
+        mov     eax,.ex2       ; calc .dby
+        sub     eax,.ex1
+        cdq
+        idiv    ebx
+        push    eax
+        mov     eax,.by2       ; calc .dex
+        sub     eax,.by1
+        cdq
+        idiv    ebx
+        push    eax
+        mov     eax,.ey2       ; calc .dey
+        sub     eax,.ey1
+        cdq
+        idiv    ebx
+        push    eax
+end if
+        mov     eax,.z2        ; calc .dz
+        sub     eax,.z1
+        cdq
+        idiv    ebx
+        push    eax
+        cmp     .x1,0         ; set correctly begin variable
+        jge     @f            ; CLIPPING ON FUNCTION
+                              ; cutting triangle exceedes screen
+        mov     ebx,.x1
+        neg     ebx
+        imul    ebx           ; eax = .dz * abs(.x1)
+        add     .z1,eax
+        mov     .x1,0
+        mov     eax,.dbx
+        imul    ebx
+        add    .bx1,eax
+        mov     eax,.dby
+        imul    ebx
+        add     .by1,eax
+        mov     eax,.dex
+        imul    ebx
+        add     .ex1,eax
+        mov     eax,.dey
+        imul    ebx
+        add     .ey1,eax
+      @@:
+        movzx   eax,word[size_x_var] ;SIZE_X  ;word[size_x_var]
+        mov     ebx,.x2
+        cmp     eax,ebx
+        jg      @f
+        mov     .x2,eax
+      @@:
+      ;  movd    mm0,eax
+      ;  movd    mm1,.x2
+      ;  pminsw  mm0,mm1
+      ;  movd    .x2,mm0
+ ;       cmp     .x2,SIZE_X  ;eax   |
+ ;       jl      @f                 |>       this dont work idk cause
+ ;       mov     .x2,SIZE_X ;eax    |
+      @@:
+      ;  movzx   eax,word[size_x_var]       ;calc memory begin in buffers
+        mov     ebx,.y
+        mul     ebx
+        mov     ebx,.x1
+        add     eax,ebx
+        mov     ebx,eax
+        lea     eax,[eax*3]
+        add     edi,eax           ; edi - screen
+        mov     esi,.z_buff       ; z-buffer filled with dd variables
+        shl     ebx,2
+        add     esi,ebx           ; esi - Z buffer
+        mov     ecx,.x2
+        sub     ecx,.x1
+        ; init current variables
+        push    dword .bx1 ;.by1 .ex1 .ey1 .z1 esi
+        push    dword .ex1
+        push    dword .by1
+        push    dword .ey1
+        push    .z1              ; current z shl CATMULL_SHIFT
+        push    esi
+if Ext >= MMX
+     pxor   mm0,mm0
+     movq   mm3,.cex   ; hi - lo -> cbx; cex
+     movq   mm4,.cey   ; hi - lo -> cby; cey
+;     movq   mm5,mm3
+;     movq   mm6,mm4
+;     psrad  mm5,ROUND
+;     psrad  mm6,ROUND
+;     movq   .ceyq,mm5
+;     movq   .cbyq,mm6
+     mov    edx,.czbuff
+else
+     cld
+end if
+     .draw:
+    ; if TEX = SHIFTING   ;bump drawing only in shifting mode
+if Ext=NON
+        mov     esi,.czbuff      ; .czbuff current address in buffer
+        mov     ebx,.cz          ; .cz - cur z position
+        cmp     ebx,dword[esi]
+else
+        mov     ebx,.cz
+        cmp     ebx,dword[edx]
+end if
+        jge     .skip
+if Ext=NON
+        mov     eax,.cby
+        mov     esi,.cbx
+        sar     eax,ROUND
+        sar     esi,ROUND
+        shl     eax,TEX_SHIFT   ;-
+        add     esi,eax
+        lea     esi,[esi*3]         ;-  ; esi - current b. texture addres
+        add     esi,.bmap
+        mov     ebx,.cex       ;.cex - current env map X
+        mov     eax,.cey       ;.cey - current  env map y
+        sar     ebx,ROUND
+        sar     eax,ROUND
+        shl     eax,TEX_SHIFT
+        add     ebx,eax
+        lea     ebx,[ebx*3]
+        add     ebx,.emap
+else
+        movq    mm5,mm4 ;.cey
+        psrad   mm5,ROUND
+        pslld   mm5,TEX_SHIFT
+        movq    mm6,mm3 ;.cex
+        psrad   mm6,ROUND
+        paddd   mm5,mm6
+        movq    mm6,mm5
+        paddd   mm5,mm5
+        paddd   mm5,mm6
+        paddd   mm5,.emap
+        movd    esi,mm5
+        psrlq   mm5,32
+        movd    ebx,mm5
+end if
+if Ext>=MMX
+        movd      mm1,[esi]
+        movd      mm2,[ebx]
+        punpcklbw mm1,mm0
+        punpcklbw mm2,mm0
+        pmullw    mm1,mm2
+        psrlw     mm1,8
+        packuswb  mm1,mm0
+        movd      [edi],mm1
+        mov       ebx,.cz
+        mov       dword[edx],ebx
+else
+        cld                     ; esi - tex e.
+        lodsb                   ; ebx - tex b.
+        mov     dl,[ebx]
+        mul     dl
+        shr     ax,8
+        stosb
+        inc     ebx
+        lodsb
+        mov     dl,[ebx]
+        mul     dl
+        shr     ax,8
+        stosb
+        inc     ebx
+        lodsb
+        mov     dl,[ebx]
+        mul     dl
+        shr     ax,8
+        stosb
+        mov     ebx,.cz
+        mov     esi,.czbuff
+        mov     dword[esi],ebx
+        jmp     .no_skip
+end if
+     .skip:
+        add     edi,3
+   if Ext = NON
+     .no_skip:
+        add     .czbuff,4
+        mov     eax,.dbx
+        add     .cbx,eax
+        mov     eax,.dby
+        add     .cby,eax
+        mov     eax,.dex
+        add     .cex,eax
+        mov     eax,.dey
+        add     .cey,eax
+    else
+        add     edx,4
+        paddd   mm3,.dex
+        paddd   mm4,.dey
+  ;      movq    mm5,mm3
+  ;      movq    mm6,mm4
+  ;      psrad   mm5,ROUND
+  ;      psrad   mm6,ROUND
+     ;   movq    .cex,mm3
+     ;   movq    .cey,mm4
+    end if
+        mov     eax,.dz
+        add     .cz,eax
+    if Ext = NON
+        dec     ecx
+        jnz     .draw
+    else
+        loop    .draw
+    end if
+  .bl_end:
+        mov     esp,ebp
+ret 56

 /programs/demos/view3ds/view3ds.asm
 ,5 → 1,5
-; application : View3ds ver. 0.071 - tiny .3ds and .asc files viewer
+; application : View3ds ver. 0.074 - tiny .3ds and .asc files viewer
 ;               with a few graphics effects demonstration.
 ; compiler    : FASM
 ; system      : KolibriOS
 ,6 → 38,9
 SSE2  =   3
 SSE3  =   4
 Ext   =   SSE3           ;Ext={ NON | MMX | SSE | SSE2 | SSE3 }
+; For now correct only SSE2 and SSE3 versions. if you have older CPU
+; use older versions of app. Probably ver 005 will be OK but it need
+; re-edit to support new Kolibri features.
 ; 0 for short names (Menuet-compatible), 1 for long names (Kolibri features)
 USE_LFN = 1  ; App is Kolibri only now.
 ,6 → 57,8
         dd     0x0              ; I_Icon
 START:    ; start of execution
+        rdtsc
+        mov    [rand_seed],ax
         cld
         push   dword (SIZE_Y shr 3) * 3
         fninit
 ,9 → 95,7
                                     ;  set point(0,0,0) in center and  calc all coords
                                     ;  to be in <-1.0,1.0>
         call   normalize_all_light_vectors
-      if Ext >= SSE3
         call   copy_lights ; to aligned float
-      end if
         call   init_triangles_normals2
         call   init_point_normals
         call   init_envmap2
 ,17 → 103,25
         call   generate_texture2
         call   init_sincos_tab
         call   do_color_buffer   ; intit color_map
+     if Ext >= SSE3
+        call   init_point_lights
+        mov    [fire_flag],0     ; proteza
+       end if
         mov    edi,bumpmap
         call   calc_bumpmap
         call   calc_bumpmap_coords   ; bump and texture mapping
         call   do_edges_list
         call   draw_window
-        ;mov     [draw_win_at_first],0
-        ;mov    eax,40         ; set events mask
-        ;mov    ebx,1100000000000000000000000100111b
-        ;int    0x40
+     if Ext > SSE2
+        mov     eax,1
+        cpuid
+        bt      ecx,0  ; is sse3 on board?
+        jc      @f
+        mov     [max_dr_flg],12
+        mov     [isSSE3],0
+       @@:
+     end if
 still:
         cmp    [edit_flag],1
         jne    @f
 ,7 → 133,16
         mov    ebx,111b
       .int:
         int    0x40
+      if Ext > SSE2
+        cmp    [ray_shd_flag],1
+        jne    @f
+        cmp    [isSSE3],1
+        jne    @f
+        mov    eax,10
+        jmp    .intt
+      end if
+       @@:
         mov     eax,23
         mov     ebx,TIMEOUT
         cmp     [speed_flag],0
 ,6 → 154,7
         mov     eax,10
       @@:
+      .intt:
         int     0x40
         cmp     eax,1           ; redraw event ?
 ,8 → 284,9
         jne     .next_m5                                ;    'grd '  1
         call    make_random_lights                      ;    'env '  2
         call    normalize_all_light_vectors             ;    'bump'  3
+        call    copy_lights
       if Ext >= SSE3
-        call    copy_lights
+        call   init_point_lights  ; for ex. ray casting
       end if
         call    do_color_buffer   ; intit color_map     ;    'tex '  4
 ,7 → 418,7
         cmp     [move_flag],0
         jne     @f
         .x_minus:
-        sub     [vect_x],10
+        sub     word[vect_x],10
         jmp     .next2
       @@:
         cmp     [move_flag],1
 ,7 → 436,7
         cmp     [move_flag],0
         jne     @f
         .x_plus:
-        add     [vect_x],10
+        add     word[vect_x],10
         jmp     .next3
       @@:
         cmp     [move_flag],1
 ,8 → 544,6
       .no_sort:
         cmp     [dr_flag],7       ; fill if 2tex and texgrd
         jge     @f
-        cmp     [catmull_flag],0  ;non fill if Catmull = off
-        je      .non_f
         cmp     [dr_flag],6       ; non fill if dots
         je      .non_f
       @@:
 ,46 → 554,74
         call     draw_dots
         jmp      .blurrr
       @@:
+      if Ext > SSE2
+        cmp     [ray_shd_flag],1  ;non fill if Catmull = off
+        jne     @f
+        cmp     [isSSE3],1
+        jne     @f
+        mov     ax,100
+        jmp     .dr
+       @@:
+      end if
+        movzx   ax,[dr_flag]
+      .dr:
         call    draw_triangles  ; draw all triangles from the list
         cmp    [edit_flag],0
         jz     .no_edit
         call   clear_vertices_index
-        call   draw_handlers
+        movzx eax,[dr_flag]
+        movzx ebx,[ray_shd_flag]
+        shl   ebx,10
+        or    eax,ebx
+         call   draw_handlers
     ;    call   edit
     .no_edit:
       .blurrr:
-        cmp  [sinus_flag],0
-        je   @f
-        call do_sinus
+        movzx eax,[dr_flag]
+        movzx ebx,[ray_shd_flag]
+        shl   ebx,10
+        or    eax,ebx
+        cmp   [sinus_flag],0
+        je    .no_sin
+        movzx eax,[dr_flag]
+        movzx ebx,[ray_shd_flag]
+        shl   ebx,10
+        or    eax,ebx
+        call  do_sinus
+      ;  jmp   .finito
+      .no_sin:
       @@:
-        cmp     [fire_flag],0
-        jne     @f
+        movzx   ecx,[fire_flag]
+        cmp     [fire_flag],1
+        je      @f
         cmp     [blur_flag],0
         je      .no_blur  ; no blur, no fire
         movzx   ecx,[blur_flag]
+      @@:
+        movzx   eax,[dr_flag]
+        movzx   ebx,[ray_shd_flag]
+        shl     ebx,10
+        or      eax,ebx
         call    blur_screen    ; blur and fire
-        jmp     .no_blur
-    @@:
-        cmp     [emboss_flag],0
-        jne     .emb           ; if emboss=true -> no fire
-        movzx   ecx,[fire_flag]
-        call    blur_screen    ; blur and fire
+     ;   jmp     .finito
     .no_blur:                  ; no blur, no fire
         cmp     [emboss_flag],0
         je      @f
-     .emb:
+        movzx eax,[dr_flag]
+        movzx ebx,[ray_shd_flag]
+        shl   ebx,10
+        or    eax,ebx
         call    do_emboss
+    .finito:
+     @@:
-      @@:
     cmp     [inc_bright_flag],0           ; increase brightness
     je      .no_inc_bright
     movzx   ebx,[inc_bright_flag]
 ,7 → 754,9
     mov     eax,7           ; put image
     mov     ebx,[screen_ptr]
     mov     ecx,[size_y_var]
-    mov    edx,[offset_y]
+    mov     edx,[offset_y]
+    cmp     [ray_shd_flag],1
+    jge     .ff
     cmp     [dr_flag],11
     jge     .ff
     int     0x40
 ,6 → 814,7
 include '3stencil.inc'
 include '3glass.inc'
 include '3glass_tex.inc'
+include '3ray_shd.inc'
 end if
 clear_vertices_index:
     mov   edi,[vertices_index_ptr]
 ,7 → 1126,7
       fldpi
       fadd      st,st
       mov       esi,[points_ptr]
-      mov       edi,tex_points
+      mov       edi,[tex_points_ptr]
       mov       ecx,[points_count_var]
       inc       ecx
 ;      cmp       [map_tex_flag],1
 ,7 → 1440,8
          mov     esp,ebp
          pop     ebp
 ret
-if Ext >= SSE3
+if   Ext >= SSE2
 init_point_normals:
 .z equ dword [ebp-8]
 .y equ dword [ebp-12]
 ,7 → 1449,6
 .point_number equ dword [ebp-28]
 .hit_faces    equ dword [ebp-32]
-        fninit
         push      ebp
         mov       ebp,esp
         sub       esp,64
 ,19 → 1489,25
         jne       .ipn_check_face
         cvtsi2ss  xmm6,.hit_faces
         movaps    xmm7,.x
         rcpss     xmm6,xmm6
         shufps    xmm6,xmm6,11000000b
         mulps     xmm7,xmm6
-        movaps    xmm6,xmm7
-        mulps     xmm6,xmm6
-        andps     xmm6,[zero_hgst_dd]
-        haddps    xmm6,xmm6
-        haddps    xmm6,xmm6
-        rsqrtps    xmm6,xmm6
-        mulps     xmm7,xmm6
         movlps    [edi],xmm7
         movhlps   xmm7,xmm7
         movss     [edi+8],xmm7
+        call      normalize_vector
+    ;    movaps    xmm6,xmm7
+    ;    mulps     xmm6,xmm6
+    ;    andps     xmm6,[zero_hgst_dd]
+    ;    haddps    xmm6,xmm6
+    ;    haddps    xmm6,xmm6
+    ;    rsqrtps    xmm6,xmm6
+    ;    mulps     xmm7,xmm6
+    ;    movlps    [edi],xmm7
+    ;    movhlps   xmm7,xmm7
+    ;    movss     [edi+8],xmm7
         add       edi,12
         inc       .point_number
         mov       edx,.point_number
 ,11 → 1633,9
         pop     ecx
         sub     ecx,1
         jnz     @b
-       ; cmp     dword[ebp],-1
-       ; jne     @b
 ret
-if Ext >= SSE3
 copy_lights: ; after normalising !
         mov      esi,lights
         mov      edi,lights_aligned
 ,8 → 1665,8
         pop      ecx
         loop     .again
 ret
-end if
 clrscr:
         mov     edi,[screen_ptr]
         movzx   ecx,word[size_x_var]
 ,7 → 1709,37
 draw_triangles:
+;  in:  eax - render draw model
+        .tri_no          equ dword[ebp-60]
+        .point_index3    equ [ebp-8]
+        .point_index2    equ [ebp-12]
+        .point_index1    equ [ebp-16]
+        .yy3             equ [ebp-18]
+        .xx3             equ [ebp-20]
+        .yy2             equ [ebp-22]
+        .xx2             equ [ebp-24]
+        .yy1             equ [ebp-26]
+        .xx1             equ [ebp-28]
+        .zz3             equ [ebp-30]
+        .zz2             equ [ebp-32]
+        .zz1             equ [ebp-34]
+        .index3x12       equ [ebp-38]
+        .index2x12       equ [ebp-42]
+        .index1x12       equ [ebp-46]
+        .temp1           equ dword[ebp-50]
+        .temp2           equ dword[ebp-54]
+        .dr_flag         equ word[ebp-56]
+        push    ebp
+        mov     ebp,esp
+        sub     esp,60
+ ;       movzx   ax,[dr_flag]
+        mov     .dr_flag,ax
         emms
       ;  update translated list  MMX required
         cmp     [vertex_edit_no],0
 ,7 → 1756,8
         movd    dword[eax],mm1
      @@:
      if Ext >= SSE3
-        cmp     [dr_flag],13
+        cmp     .dr_flag,13
         jnge    .no_stencil
         mov     esi,[triangles_ptr]
         mov     ecx,[triangles_count_var]
 ,124 → 1820,57
         je      .draw_smooth_line
         mov esi,[triangles_ptr]
-        mov ecx,[triangles_count_var]
+        xor ecx,ecx  ;mov ecx,[triangles_count_var]
     .again_dts:
+       ; push    ebp
+        push    esi
         push    ecx
-        mov     ebp,[points_translated_ptr]
-      if Ext >= SSE2
-        mov     eax,dword[esi]
-        mov     [point_index1],eax
-        lea     eax,[eax*3]
-        add     eax,eax
-        push    ebp
-        add     ebp,eax
-        mov     eax,[ebp]
-   ;     cmp     [vertex_edit_no],0
-   ;     jne     @f
-   ;
-   ;   @@:
-        mov     dword[xx1],eax
-        mov     eax,[ebp+4]
-        mov     [zz1],ax
+        mov     .tri_no,ecx
-        pop     ebp
+        mov     eax,[esi]
+        mov     ebx,[esi+4]
+        mov     ecx,[esi+8]
+        mov     .point_index1,eax
+        mov     .point_index2,ebx
+        mov     .point_index3,ecx
+        imul    eax,[i12]
+        imul    ebx,[i12]
+        imul    ecx,[i12]
+        mov     .index1x12,eax
+        mov     .index2x12,ebx
+        mov     .index3x12,ecx
-        mov     eax,dword[esi+4]
-        mov     [point_index2],eax
-        lea     eax,[eax*3]
-        add     eax,eax
-        push    ebp
-        add     ebp,eax
-        mov     eax,[ebp]
-        mov     dword[xx2],eax
-        mov     eax,[ebp+4]
-        mov     [zz2],ax
-        pop     ebp
+        shr     eax,1
+        shr     ebx,1
+        shr     ecx,1
+        add     eax,[points_translated_ptr]
+        add     ebx,[points_translated_ptr]
+        add     ecx,[points_translated_ptr]
+        push    word[eax+4]
+        push    word[ebx+4]
+        push    word[ecx+4]
+        pop     word .zz3
+        pop     word .zz2
+        pop     word .zz1
+        mov     eax,[eax]
+        mov     ebx,[ebx]
+        mov     ecx,[ecx]
+        ror     eax,16
+        ror     ebx,16
+        ror     ecx,16
+        mov     .xx1,eax
+        mov     .xx2,ebx
+        mov     .xx3,ecx
-        mov     eax,dword[esi+8]        ; xyz3 = [ebp+[esi+4]*6]
-        mov     [point_index3],eax
-        lea     eax,[eax*3]
-        add     eax,eax
-    ;    push    ebp
-        add     ebp,eax
-        mov     eax,[ebp]
-        mov     dword[xx3],eax
-        mov     eax,[ebp+4]
-        mov     [zz3],ax
-      else
-        movq    mm0,[esi]           ; don't know MMX
-        mov     qword[point_index1],mm0
-       ; shr     eax,16
-       ; mov     [point_index2],ax
-        mov     eax,dword[esi+8]
-        mov     [point_index3],eax
-        movdqu  xmm0,[esi]
-        paddd   xmm0,xmm0
-        movdqa  xmm1,xmm0
-        paddd   xmm0,xmm0
-        paddd   xmm0,xmm1
-        movd    eax,xmm0
-        psrldq  xmm0,4
-        movd    ebx,xmm0
-        psrldq  xmm0,4
-        movd    ecx,xmm0
-        and     eax,0FFFFh
-        and     ebx,0FFFFh
-        and     ecx,0FFFFh
-        movq    mm0,[ebp+eax]
-        movq    mm1,[ebp+ebx]
-        movq    mm2,[ebp+ecx]
-        movq    qword[xx1],mm0
-        movq    qword[xx2],mm1
-        movq    qword[xx3],mm2
-;        emms
-      end if  ; *********************************
- if 0
-        cmp     [vertex_edit_no],0
-        jne     .no_edit
-        mov     ax,[vertex_edit_no]
-        dec     ax
-        cmp     ax,[point_index1]
-        jne     @f
-        movd    mm0,[edit_start_x]
-        psubw   mm0,[edit_end_x]
-        movd    mm1,dword[xx1]
-        paddw   mm1,mm0
-        movd    dword[xx1],mm1
-        jmp     .no_edit
-       @@:
-        cmp     ax,[point_index2]
-        jne     @f
-        movd    mm0,[edit_start_x]
-        psubw   mm0,[edit_end_x]
-        movd    mm1,dword[xx2]
-        paddw   mm1,mm0
-        movd    dword[xx2],mm1
-        jmp     .no_edit
-       @@:
-        cmp     ax,[point_index3]
-        jne     @f
-        movd    mm0,[edit_start_x]
-        psubw   mm0,[edit_end_x]
-        movd    mm1,dword[xx3]
-        paddw   mm1,mm0
-        movd    dword[xx3],mm1
-        jmp     .no_edit
-       @@:
-    .no_edit:
-end if
-        push esi                          ;
+      ;  push    esi
         fninit                            ; DO culling AT FIRST
         cmp     [culling_flag],1          ; (if culling_flag = 1)
         jne     .no_culling
-        mov     esi,point_index1          ; *********************************
+        lea     esi,.point_index1          ; *********************************
         mov     ecx,3                     ;
       @@:
         mov     eax,dword[esi]
 ,76 → 1877,77
         lea     eax,[eax*3]
         shl     eax,2
         add     eax,[points_normals_rot_ptr]
-;        lea     eax,[eax+point_normals_rotated]
-        fld     dword[eax+8]             ; *****************************
-        ftst                             ; CHECKING OF Z COOFICIENT OF
-        fstsw   ax                       ; NORMAL VECTOR
-        sahf
-        jb      @f
-        ffree   st
+        mov     eax,[eax+8]
+        bt      eax,31
+        jc      @f
+                       ; *****************************
+                       ; CHECKING OF Z COOFICIENT OF
+                       ; NORMAL VECTOR
+        add     esi,4
         loop    @b
         jmp     .end_draw   ; non visable
       @@:
-        ffree   st  ;is visable
       .no_culling:
-        cmp     [dr_flag],0               ; draw type flag
+        cmp     .dr_flag,0               ; draw type flag
         je      .flat_draw
-        cmp     [dr_flag],2
+        cmp     .dr_flag,2
         je      .env_mapping
-        cmp     [dr_flag],3
+        cmp     .dr_flag,3
         je      .bump_mapping
-        cmp     [dr_flag],4
+        cmp     .dr_flag,4
         je      .tex_mapping
-        cmp     [dr_flag],5
+        cmp     .dr_flag,5
         je      .rainbow
-        cmp     [dr_flag],7
+        cmp     .dr_flag,7
         je      .grd_tex
-        cmp     [dr_flag],8
+        cmp     .dr_flag,8
         je      .two_tex
-        cmp     [dr_flag],9
+        cmp     .dr_flag,9
         je      .bump_tex
-        cmp     [dr_flag],10
+        cmp     .dr_flag,10
         je      .cubic_env_mapping
-        cmp     [dr_flag],11
+        cmp     .dr_flag,11
         je      .draw_smooth_line
       if Ext >= SSE3
-        cmp     [dr_flag],12
+        cmp     .dr_flag,12
         je      .r_phg
-        cmp     [dr_flag],13
+        cmp     .dr_flag,13
         je      .glass
-        cmp     [dr_flag],14
+        cmp     .dr_flag,14
         je      .glass_tex
-     end if                                 ; ****************
-        mov     esi,point_index3      ; do Gouraud shading
+        cmp     .dr_flag,100
+        je      .ray_shd
+     end if
+        push    ebp                    ; ****************
+        lea     esi,.index3x12      ; do Gouraud shading
+        lea     edi,.zz3
         mov     ecx,3
       .again_grd_draw:
         mov     eax,dword[esi]
-        shl     eax,2
-        lea     eax,[eax*3]
         add     eax,[points_normals_rot_ptr]
         ; texture x=(rotated point normal -> x * 255)+255
         fld     dword[eax]       ; x cooficient of normal vector
         fimul   [correct_tex]
         fiadd   [correct_tex]
-        fistp   [temp1]
+        fistp   .temp1
         ; texture y=(rotated point normal -> y * 255)+255
         fld     dword[eax+4]      ; y cooficient
         fimul   [correct_tex]
         fiadd   [correct_tex]
-        fistp   [temp2]
+        fistp   .temp2
-        mov      eax,[temp2]
-        mov      ebx,[temp1]
+        mov      eax,.temp2
+        mov      ebx,.temp1
         and      ebx,0xfffffff
         shl      eax,TEX_SHIFT
         add      eax,ebx
         lea      eax,[eax*3+color_map]
         mov      eax,dword[eax]
-     ;   cmp     [catmull_flag],1      ; put on stack z coordinate if necessary
-     ;   jne      @f
-        lea      edx,[ecx*3]
-        push     word[edx*2+xx1-2]    ; zz1 ,2 ,3
-     ; @@:
+        push     word[edi]    ; zz1 ,2 ,3
         ror      eax,16               ; eax -0xxxrrggbb -> 0xggbbxxrr
         xor      ah,ah
         push     ax         ;r
 ,98 → 1958,55
         push     ax         ;b
         sub      esi,4
+        sub      edi,2
         dec      cx
         jnz      .again_grd_draw
         jmp      .both_draw
-   ;     movzx   edi,[point_index3]   ;gouraud shading according to light vector
-   ;     lea     edi,[edi*3]
-   ;     lea     edi,[4*edi+point_normals_rotated] ; edi - normal
-   ;     mov     esi,light_vector
-   ;     call    dot_product
-   ;     fabs
-   ;     fimul   [orginal_color_r]
-   ;     fistp   [temp_col]
-   ;     and     [temp_col],0x00ff
-   ;     push    [temp_col]
-   ;     push    [temp_col]
-   ;     push    [temp_col]
+   .rainbow:
+        push     ebp
+        push     word .zz3
-   ;     movzx   edi,[point_index2]
-   ;     lea     edi,[edi*3]
-   ;     lea     edi,[4*edi+point_normals_rotated] ; edi - normal
-   ;     mov     esi,light_vector
-   ;     call    dot_product
-   ;     fabs
-   ;     fimul   [orginal_color_r]
-   ;     fistp    [temp_col]
-   ;     and     [temp_col],0x00ff
-   ;     push    [temp_col]
-   ;     push    [temp_col]
-   ;     push    [temp_col]
-   ;     movzx   edi,[point_index1]
-   ;     lea     edi,[edi*3]
-   ;     lea     edi,[4*edi+point_normals_rotated] ; edi - normal
-   ;     mov     esi,light_vector
-   ;     call    dot_product
-   ;     fabs
-   ;     fimul   [orginal_color_r]
-   ;     fistp   [temp_col]
-   ;     and     [temp_col],0x00ff
-   ;     push    [temp_col]
-   ;     push    [temp_col]
-   ;     push    [temp_col]
-   .rainbow:
-       ; cmp     [catmull_flag],1      ; put on stack z coordinate if necessary
-       ; jne      @f
-        push     [zz3]
-      @@:
-        mov      eax,dword[yy3]
+        mov      eax, .xx3
+        ror      eax,16
         mov      ebx,0x00ff00ff
         and      eax,ebx
         push     eax
         neg      al
         push     ax
-        push     [zz2]
+        push     word .zz2
-        mov      eax,dword[yy2]
+        mov      eax, .xx2
+        ror      eax,16
         and      eax,ebx
         push     eax
         neg      al
         push     ax
-        push     [zz1]
+        push     word .zz1
-        mov      eax,dword[yy1]
+        mov      eax, .xx1
+        ror      eax,16
         and      eax,ebx
         push     eax
         neg      al
         push     ax
     .both_draw:
-        mov     eax,dword[xx1]
-        ror     eax,16
-        mov     ebx,dword[xx2]
-        ror     ebx,16
-        mov     ecx,dword[xx3]
-        ror     ecx,16
+        mov     eax, .xx1
+        mov     ebx, .xx2
+        mov     ecx, .xx3
         mov     edi,[screen_ptr]
         mov     esi,[Zbuffer_ptr]
         call    gouraud_triangle_z
+        pop     ebp
         jmp     .end_draw
      .flat_draw:                     ;**************************
         fninit                             ; FLAT DRAWING
-        mov     eax,[point_index1]
-        mov     ebx,[point_index2]
-        mov     ecx,[point_index3]
-        shl     eax,2
-        shl     ebx,2
-        shl     ecx,2
-        lea     eax,[eax*3]  ;+point_normals_rotated]
+        mov     eax,.index1x12
+        mov     ebx,.index2x12
+        mov     ecx,.index3x12
         add     eax,[points_normals_rot_ptr]
-        lea     ebx,[ebx*3]  ;+point_normals_rotated]
         add     ebx,[points_normals_rot_ptr]
-        lea     ecx,[ecx*3]  ;+point_normals_rotated]
         add     ecx,[points_normals_rot_ptr]
         fld     dword[eax]      ; x cooficient of normal vector
         fadd    dword[ebx]
 ,7 → 2014,7
         fidiv   [i3]
         fimul   [correct_tex]
         fiadd   [correct_tex]
-        fistp   [temp1]  ;dword[esp-4]    ; x temp variables
+        fistp   .temp1  ;dword[esp-4]    ; x temp variables
         fld     dword[eax+4]    ; y cooficient of normal vector
         fadd    dword[ebx+4]
         fadd    dword[ecx+4]
 ,12 → 2021,12
         fidiv   [i3]
         fimul   [correct_tex]
         fiadd   [correct_tex]
-        fistp   [temp2]  ;dword[esp-8]   ;  y
-        mov     edx,[temp2] ;dword[esp-8]
+        fistp   .temp2  ;dword[esp-8]   ;  y
+        mov     edx,.temp2 ;dword[esp-8]
         and     edx,0xfffffff
-        and     [temp1],0xfffffff
+        and     .temp1,0xfffffff
         shl     edx,TEX_SHIFT
-        add     edx,[temp1]  ;dword[esp-4]
+        add     edx,.temp1  ;dword[esp-4]
         lea     eax,[3*edx]
         add     eax,color_map
 ,34 → 2048,32
      ;   shl     eax,8
      ;   mov     edx,eax
-        mov     eax,dword[xx1]
-        ror     eax,16
-        mov     ebx,dword[xx2]
-        ror     ebx,16
-        mov     ecx,dword[xx3]
-        ror     ecx,16
+        mov     eax,dword .xx1
+        mov     ebx,dword .xx2
+        mov     ecx,dword .xx3
         mov     edi,[screen_ptr]
         mov     esi,[Zbuffer_ptr]
-        push    word[zz3]
-        push    word[zz2]
-        push    word[zz1]
+        push    ebp
+        push    word .zz3
+        push    word .zz2
+        push    word .zz1
         call    flat_triangle_z
+        pop     ebp
         jmp     .end_draw
       .env_mapping:
-        push    [zz3]
-        push    [zz2]
-        push    [zz1]
+        push    ebp
+        push    word .zz3
+        push    word .zz2
+        push    word .zz1
-        mov     esi,point_index1
+        lea     esi, .index1x12
         sub     esp,12
         mov     edi,esp
         mov     ecx,3
       @@:
         mov     eax,dword[esi]
-        lea     eax,[eax*3]
-        shl     eax,2
         add     eax,[points_normals_rot_ptr]       ;point_normals_rotated
         ; texture x=(rotated point normal -> x * 255)+255
         fld     dword[eax]
 ,33 → 2090,29
         add     esi,4
         loop    @b
-        mov     eax,dword[xx1]
-        ror     eax,16
-        mov     ebx,dword[xx2]
-        ror     ebx,16
-        mov     ecx,dword[xx3]
-        ror     ecx,16
+        mov     eax, .xx1
+        mov     ebx,dword .xx2
+        mov     ecx,dword .xx3
         mov     edi,[screen_ptr]
         mov     esi,envmap
         mov     edx,[Zbuffer_ptr]
         call    tex_triangle_z
+        pop     ebp
         jmp     .end_draw
 ;+++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
      .cubic_env_mapping:
-        push    [zz3]
-        push    [zz2]
-        push    [zz1]
+        push    ebp
+        push    word .zz3
+        push    word .zz2
+        push    word .zz1
-        mov     esi,point_index1
+        lea     esi,.index1x12
         sub     esp,12
         mov     edi,esp
         mov     ecx,3
       @@:
         mov     eax,dword[esi]
-        lea     eax,[eax*3]
-        shl     eax,2
         add     eax,[points_normals_rot_ptr]
         fld     dword[eax]
 ,37 → 2142,32
         add     esi,4
         loop    @b
-        mov     eax,dword[xx1]
-        ror     eax,16
-        mov     ebx,dword[xx2]
-        ror     ebx,16
-        mov     ecx,dword[xx3]
-        ror     ecx,16
+        mov     eax, .xx1
+        mov     ebx, .xx2
+        mov     ecx, .xx3
         mov     edi,[screen_ptr]
         mov     esi,envmap_cub
         mov     edx,[Zbuffer_ptr]
         call    tex_triangle_z
+        pop     ebp
         jmp     .end_draw
 ;+++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
       .bump_mapping:
+        push    ebp
         push    [Zbuffer_ptr]
-        push    [zz3]
-        push    [zz2]
-        push    [zz1]
+        push    word .zz3
+        push    word .zz2
+        push    word .zz1
-        mov     esi,point_index1
+        lea     esi,.index1x12
         sub     esp,12
         mov     edi,esp
         mov     ecx,3
       @@:
         mov     eax,dword[esi]
-        lea     eax,[eax*3]
-        shl     eax,2
         add     eax,[points_normals_rot_ptr]  ;point_normals_rotated
         ; texture x=(rotated point normal -> x * 255)+255
         fld     dword[eax]
 ,70 → 2184,58
         add     esi,4
         loop    @b
-        mov    esi,[point_index3]      ; bump map coords
+        mov    esi, .point_index3      ; bump map coords
         shl    esi,2
-        add    esi,tex_points
+        add    esi,[tex_points_ptr]
         push   dword[esi]
-        mov    esi,[point_index2]
+        mov    esi, .point_index2
         shl    esi,2
-        add    esi,tex_points
-;       lea    esi,[esi*3]
-;       lea    esi,[points+2+esi*2]
+        add    esi,[tex_points_ptr]
         push   dword[esi]
-  ;     push   dword[xx2]
-        mov    esi,[point_index1]
+        mov    esi, .point_index1
         shl    esi,2
-        add    esi,tex_points
-;       lea     esi,[esi*3]
-;       lea     esi,[points+2+esi*2]
+        add    esi,[tex_points_ptr]
         push   dword[esi]
-   ;    push     dword[xx1]
-        mov     eax,dword[xx1]
-        ror     eax,16
-        mov     ebx,dword[xx2]
-        ror     ebx,16
-        mov     ecx,dword[xx3]
-        ror     ecx,16
+        mov     eax,dword .xx1
+        mov     ebx,dword .xx2
+        mov     ecx,dword .xx3
         mov     edi,[screen_ptr]
         mov     esi,envmap
         mov     edx,bumpmap            ;BUMP_MAPPING
         call    bump_triangle_z
+        pop     ebp
         jmp     .end_draw
       .tex_mapping:
-        push    [zz3]
-        push    [zz2]
-        push    [zz1]
+        push   ebp
+        push   word .zz3
+        push   word .zz2
+        push   word .zz1
    ;   @@:
-        mov    esi,[point_index3]      ; tex map coords
+        mov    esi, .point_index3      ; tex map coords
         shl    esi,2
-        add    esi,tex_points
+        add    esi,[tex_points_ptr]
         push   dword[esi]
-        mov    esi,[point_index2]
+        mov    esi, .point_index2
         shl    esi,2
-        add    esi,tex_points
+        add    esi,[tex_points_ptr]
         push   dword[esi]
-        mov    esi,[point_index1]
+        mov    esi, .point_index1
         shl    esi,2
-        add    esi,tex_points
+        add    esi,[tex_points_ptr]
         push   dword[esi]
-        mov     eax,dword[xx1]
-        ror     eax,16
-        mov     ebx,dword[xx2]
-        ror     ebx,16
-        mov     ecx,dword[xx3]
-        ror     ecx,16
+        mov     eax,dword .xx1
+        mov     ebx,dword .xx2
+        mov     ecx,dword .xx3
         mov     edi,[screen_ptr]
         mov     esi,texmap
-         mov     edx,[Zbuffer_ptr]
+        mov     edx,[Zbuffer_ptr]
         call    tex_triangle_z
+        pop     ebp
         jmp     .end_draw
 ;      .ray:
 ;        grd_triangle according to points index
 ,49 → 2273,43
      .grd_tex:            ; smooth shading + texture
          push   ebp
-         mov    ebp,esp
-         sub    esp,4
-         push   ebp
-         mov    esi,[point_index3]      ; tex map coords
+         mov    esi, .point_index3      ; tex map coords
          shl    esi,2
-         add    esi,tex_points
+         add    esi,[tex_points_ptr]
          push   dword[esi]              ; texture coords as first
-         mov    esi,[point_index2]      ; group of parameters
+         mov    esi, .point_index2      ; group of parameters
          shl    esi,2
-         add    esi,tex_points
+         add    esi,[tex_points_ptr]
          push   dword[esi]
-         mov    esi,[point_index1]
+         mov    esi, .point_index1
          shl    esi,2
-         add    esi,tex_points
+         add    esi,[tex_points_ptr]
          push   dword[esi]
-         mov     esi,point_index3
+         lea     esi, .index3x12
+         lea     edi, .zz3
          mov     ecx,3
+       .aagain_grd_draw:
-      .aagain_grd_draw:
-        lea      edx,[ecx*3]
-        push     word[edx*2+xx1-2]    ; zz1 ,2 ,3
+        push     word[edi]    ; zz1 ,2 ,3
         fninit
         mov     eax,dword[esi]
-        shl     eax,2
-        lea     eax,[eax*3] ;+point_normals_rotated]
         add     eax,[points_normals_rot_ptr]
         ; texture x=(rotated point normal -> x * 255)+255
         fld     dword[eax]       ; x cooficient of normal vector
         fimul   [correct_tex]
         fiadd   [correct_tex]
-        fistp   [temp1]  ;word[ebp-2]
+        fistp   .temp1  ;word[ebp-2]
         ; texture y=(rotated point normal -> y * 255)+255
         fld     dword[eax+4]      ; y cooficient
         fimul   [correct_tex]
         fiadd   [correct_tex]
-        fistp   [temp2]  ;word[ebp-4]
+        fistp   .temp2  ;word[ebp-4]
-        mov      eax,[temp2]   ;word[ebp-4]
-        mov      ebx,[temp1]   ;word[ebp-2]
-        and      ebx,0xfffffff ; some onjects need thid 'and'
+        mov      eax,.temp2
+        mov      ebx,.temp1
+        and      ebx,0xfffffff ; some onjects need this 'and'
         shl      eax,TEX_SHIFT
         add      eax,ebx
         lea      eax,[eax*3]
 ,17 → 2324,14
         push     ax         ;g
         shr      eax,24
         push     ax         ;b
+        sub      edi,2
         sub      esi,4
         dec      cx
         jnz      .aagain_grd_draw
-        mov     eax,dword[xx1]
-        ror     eax,16
-        mov     ebx,dword[xx2]
-        ror     ebx,16
-        mov     ecx,dword[xx3]
-        ror     ecx,16
+        mov     eax, .xx1
+        mov     ebx, .xx2
+        mov     ecx, .xx3
         mov     edi,[screen_ptr]
         mov     edx,texmap
         mov     esi,[Zbuffer_ptr]
 ,31 → 2339,30
         call    tex_plus_grd_triangle
         pop     ebp
-        mov     esp,ebp
-        pop     ebp
         jmp     .end_draw
       .two_tex:
+        push     ebp
         push    [Zbuffer_ptr]
-        push    word[zz3]
-        push    word[zz2]
-        push    word[zz1]
+        push    word .zz3
+        push    word .zz2
+        push    word .zz1
-        mov    esi,[point_index3]      ; tex map coords
+        mov    esi, .point_index3      ; tex map coords
         shl    esi,2
-        add    esi,tex_points
+        add    esi,[tex_points_ptr]
         push   dword[esi]
-        mov    esi,[point_index2]
+        mov    esi, .point_index2
         shl    esi,2
-        add    esi,tex_points
+        add    esi,[tex_points_ptr]
         push   dword[esi]
-        mov    esi,[point_index1]
+        mov    esi, .point_index1
         shl    esi,2
-        add    esi,tex_points
+        add    esi,[tex_points_ptr]
         push   dword[esi]
-        mov     esi,point_index1     ; env coords
+        lea     esi, .point_index1     ; env coords
         sub     esp,12
         mov     edi,esp
         mov     ecx,3
 ,50 → 2387,46
         add     esi,4
         loop    @b
-        mov     eax,dword[xx1]
-        ror     eax,16
-        mov     ebx,dword[xx2]
-        ror     ebx,16
-        mov     ecx,dword[xx3]
-        ror     ecx,16
+        mov     eax, .xx1
+        mov     ebx, .xx2
+        mov     ecx, .xx3
         mov     edi,[screen_ptr]
         mov     esi,texmap
         mov     edx,envmap
         call    two_tex_triangle_z
+        pop     ebp
         jmp     .end_draw
    .bump_tex:
-        mov    esi,[point_index3]      ; tex map coords
+        push   ebp
+        mov    esi, .point_index3      ; tex map coords
         shl    esi,2
-        add    esi,tex_points
+        add    esi,[tex_points_ptr]
         push   dword[esi]
-        mov    esi,[point_index2]
+        mov    esi, .point_index2
         shl    esi,2
-        add    esi,tex_points
+        add    esi,[tex_points_ptr]
         push   dword[esi]
-        mov    esi,[point_index1]
+        mov    esi, .point_index1
         shl    esi,2
-        add    esi,tex_points
+        add    esi,[tex_points_ptr]
         push   dword[esi]
         push  dword texmap
         push  [Zbuffer_ptr]
-        xor   edi,edi
-        push    word[zz3]
-        push    word[zz2]
-        push    word[zz1]
+        push    word .zz3
+        push    word .zz2
+        push    word .zz1
-        mov     esi,point_index1     ; env coords
+        lea     esi, .index1x12     ; env coords
         sub     esp,12
         mov     edi,esp
         mov     ecx,3
       @@:
         mov     eax,dword[esi]
-        lea     eax,[eax*3]
-        shl     eax,2
         add     eax,[points_normals_rot_ptr]
         ; texture x=(rotated point normal -> x * 255)+255
         fld     dword[eax]
 ,40 → 2443,28
         add     esi,4
         loop    @b
-;        push  dword 1 shl 16 + 1  ; emap coords
-;        push  dword 127 shl 16 + 1
-;        push  dword 127 shl 16 + 127
-        mov    esi,[point_index3]      ; bump map coords
+        mov    esi, .point_index3      ; bump map coords
         shl    esi,2
-        add    esi,tex_points
+        add    esi,[tex_points_ptr]
         push   dword[esi]
-        mov    esi,[point_index2]
+        mov    esi, .point_index2
         shl    esi,2
-        add    esi,tex_points
+        add    esi,[tex_points_ptr]
         push   dword[esi]
-        mov    esi,[point_index1]
+        mov    esi, .point_index1
         shl    esi,2
-        add    esi,tex_points
+        add    esi,[tex_points_ptr]
         push   dword[esi]
-;        push  dword 1 shl 16 + 127
-;        push  dword 127 shl 16 + 127
-;        push  dword 1 shl 16 + 1  ; bump coords
-        mov     eax,dword[xx1]
-        ror     eax,16
-        mov     ebx,dword[xx2]
-        ror     ebx,16
-        mov     ecx,dword[xx3]
-        ror     ecx,16
+        mov     eax,dword .xx1
+        mov     ebx,dword .xx2
+        mov     ecx,dword .xx3
         mov     edi,[screen_ptr]
         mov     esi,envmap
         mov     edx,bumpmap
         call bump_tex_triangle_z
+        pop     ebp
         jmp     .end_draw
 ,12 → 2477,9
         pshufd    xmm5,xmm5,01110011b
-        mov     eax,[point_index1]
-        mov     ebx,[point_index2]
-        mov     ecx,[point_index3]
-        imul    eax,[i12]
-        imul    ebx,[i12]
-        imul    ecx,[i12]
+        mov     eax, .index1x12
+        mov     ebx, .index2x12
+        mov     ecx, .index3x12
         add     eax,[points_normals_rot_ptr]
         add     ebx,[points_normals_rot_ptr]
         add     ecx,[points_normals_rot_ptr]
 ,12 → 2491,9
         andps   xmm2,[zero_hgst_dd]
         xorps   xmm3,xmm3
-        mov     eax,[point_index1]
-        mov     ebx,[point_index2]
-        mov     ecx,[point_index3]
-        imul    eax,[i12]
-        imul    ebx,[i12]
-        imul    ecx,[i12]
+        mov     eax, .index1x12
+        mov     ebx, .index2x12
+        mov     ecx, .index3x12
         add     eax,[points_rotated_ptr]
         add     ebx,[points_rotated_ptr]
         add     ecx,[points_rotated_ptr]
 ,12 → 2506,9
-        mov     eax,dword[xx1]
-        ror     eax,16
-        mov     ebx,dword[xx2]
-        ror     ebx,16
-        mov     ecx,dword[xx3]
-        ror     ecx,16
+        mov     eax,dword .xx1
+        mov     ebx,dword .xx2
+        mov     ecx,dword .xx3
         mov     edi,[screen_ptr]
         mov     esi,[Zbuffer_ptr]
 ,12 → 2522,9
         pshufd    xmm5,xmm5,01110011b
-        mov     eax,[point_index1]
-        mov     ebx ,[point_index2]
-        mov     ecx,[point_index3]
-        imul    eax,[i12]
-        imul    ebx,[i12]
-        imul    ecx,[i12]
+        mov     eax, .index1x12
+        mov     ebx, .index2x12
+        mov     ecx, .index3x12
         add     eax,[points_normals_rot_ptr]
         add     ebx,[points_normals_rot_ptr]
         add     ecx,[points_normals_rot_ptr]
 ,12 → 2536,9
         andps   xmm2,[zero_hgst_dd]
         xorps   xmm3,xmm3
-        mov     eax,[point_index1]
-        mov     ebx,[point_index2]
-        mov     ecx,[point_index3]
-        imul    eax,[i12]
-        imul    ebx,[i12]
-        imul    ecx,[i12]
+        mov     eax, .index1x12
+        mov     ebx, .index2x12
+        mov     ecx, .index3x12
         add     eax,[points_rotated_ptr]
         add     ebx,[points_rotated_ptr]
         add     ecx,[points_rotated_ptr]
 ,12 → 2551,9
-        mov     eax,dword[xx1]
-        ror     eax,16
-        mov     ebx,dword[xx2]
-        ror     ebx,16
-        mov     ecx,dword[xx3]
-        ror     ecx,16
+        mov     eax, .xx1
+        mov     ebx, .xx2
+        mov     ecx, .xx3
         mov     edi,[screen_ptr]
         mov     edx,[Zbuffer_ptr]
         mov     esi,[Zbuffer_ptr]
 ,12 → 2567,9
         punpcklwd xmm5,[the_zero]
         pshufd    xmm5,xmm5,01110011b
-        mov     eax,[point_index1]
-        mov     ebx,[point_index2]
-        mov     ecx,[point_index3]
-        imul    eax,[i12]
-        imul    ebx,[i12]
-        imul    ecx,[i12]
+        mov     eax, .index1x12
+        mov     ebx, .index2x12
+        mov     ecx, .index3x12
         add     eax,[points_normals_rot_ptr]
         add     ebx,[points_normals_rot_ptr]
         add     ecx,[points_normals_rot_ptr]
 ,12 → 2581,9
         andps   xmm2,[zero_hgst_dd]
         xorps   xmm3,xmm3
-        mov     eax,[point_index1]
-        mov     ebx,[point_index2]
-        mov     ecx,[point_index3]
-        imul    eax,[i12]
-        imul    ebx,[i12]
-        imul    ecx,[i12]
+        mov     eax, .index1x12
+        mov     ebx, .index2x12
+        mov     ecx, .index3x12
         add     eax,[points_rotated_ptr]
         add     ebx,[points_rotated_ptr]
         add     ecx,[points_rotated_ptr]
 ,17 → 2594,17
         add     esp,12
         andps   xmm4,[zero_hgst_dd]
-        mov    esi,[point_index3]      ; tex map coords
+        mov    esi,.point_index3      ; tex map coords
         shl    esi,2
-        add    esi,tex_points
+        add    esi,[tex_points_ptr]
         push   dword[esi]
-        mov    esi,[point_index2]
+        mov    esi,.point_index2
         shl    esi,2
-        add    esi,tex_points
+        add    esi,[tex_points_ptr]
         push   dword[esi]
-        mov    esi,[point_index1]
+        mov    esi,.point_index1
         shl    esi,2
-        add    esi,tex_points
+        add    esi,[tex_points_ptr]
         push   dword[esi]
         movups xmm6,[esp]
         add    esp,12
 ,31 → 2618,84
         por    xmm6,xmm7
-        mov     eax,dword[xx1]
-        ror     eax,16
-        mov     ebx,dword[xx2]
-        ror     ebx,16
-        mov     ecx,dword[xx3]
-        ror     ecx,16
+        mov     eax,dword .xx1
+        mov     ebx,dword .xx2
+        mov     ecx,dword .xx3
         mov     edx,texmap
         mov     edi,[screen_ptr]
         mov     esi,[Zbuffer_ptr]
         call    glass_tex_tri
+        jmp     .end_draw
+   .ray_shd:
+        emms
+        movd      xmm5,[size_y_var]
+        punpcklwd xmm5,[the_zero]
+        pshufd    xmm5,xmm5,01110011b
+        mov     eax, .index1x12
+        mov     ebx, .index2x12
+        mov     ecx, .index3x12
+        add     eax,[points_normals_rot_ptr]
+        add     ebx,[points_normals_rot_ptr]
+        add     ecx,[points_normals_rot_ptr]
+        movups  xmm0,[eax]
+        movups  xmm1,[ebx]
+        movups  xmm2,[ecx]
+        andps   xmm0,[zero_hgst_dd]
+        andps   xmm1,[zero_hgst_dd]
+        andps   xmm2,[zero_hgst_dd]
+        xorps   xmm3,xmm3
+   ;     mov     ebx,.tri_no
+   ;     cmp     ebx,0
+   ;     je      @f
+   ;     int3
+   ;   @@:
+        mov     eax, .index1x12
+        mov     ebx, .index2x12
+        mov     ecx, .index3x12
+        add     eax,[points_rotated_ptr]
+        add     ebx,[points_rotated_ptr]
+        add     ecx,[points_rotated_ptr]
+        push    dword[ecx+8]
+        push    dword[ebx+8]
+        push    dword[eax+8]
+        movups  xmm4,[esp]
+        add     esp,12
+        andps   xmm4,[zero_hgst_dd]
+        movd    mm7,.tri_no
+   ;     mm7 - intialised
+        mov     eax,dword .xx1
+        mov     ebx,dword .xx2
+        mov     ecx,dword .xx3
+        mov     edx,texmap
+        mov     edi,[screen_ptr]
+        mov     esi,[Zbuffer_ptr]
+        call    ray_shad
 end if
       .end_draw:
+    ;    pop     ebp
+        pop     ecx
         pop     esi
         add     esi,12
+        inc     ecx
+        cmp     ecx,[triangles_count_var]
+        jnz     .again_dts
-        pop     ecx
-        dec     ecx
+        jmp     .eend
-        jnz     .again_dts
-ret
    .draw_smooth_line:
 ,7 → 2746,8
         sub       esp,16
         movups    [esp],xmm1
         add       esi,4
-        loop      .aga_n
+        dec       ecx
+        jnz      .aga_n
         movups    xmm0,[esp]
         movups    xmm1,[esp+16]
 ,11 → 2765,17
         movhps  xmm7,[edx]
         pshufd  xmm7,xmm7,11101000b
         movdqa  xmm6,xmm7
+        movdqa  xmm3,xmm7
+        movdqa  xmm4,xmm7
         movd    xmm5,[size_y_var]
         pshuflw xmm5,xmm5,00010001b
+        pcmpeqw xmm3,xmm5
+        pcmpeqw xmm4,[the_zero]
         pcmpgtw xmm7,xmm5
         pcmpgtw xmm6,[the_zero]
         pxor    xmm7,xmm6
+        pxor    xmm3,xmm4
+        pxor    xmm7,xmm3
         pmovmskb eax,xmm7
         cmp     al,-1
         jnz      .skp
 ,15 → 2815,25
         cmp     ecx,[edges_count]
         jnz     .again_s_line
- ret
+   .eend:
+        add      esp,60
+        pop      ebp
+ret
 draw_handlers:
+       ;  in eax - render model
        push  ebp
        mov   ebp,esp
 ,14 → 2840,15
        .counter  equ ebp-16
        .xres3m18 equ ebp-8
        .xres2m12 equ ebp-12
+       .dr_model equ dword[ebp-4]
      ; init counter
        sub   esp,12
        push  dword 0
+       mov   .dr_model,eax
        movzx eax,word[size_x_var]
-       cmp    [dr_flag],12
+       cmp    .dr_model,12
        jge    @f
        lea   ebx,[eax*3]
        sub   ebx,18
 ,7 → 2906,7
        add   eax,ebx
        push   eax
        lea   edi,[eax*3]
-       cmp    [dr_flag],12
+       cmp    .dr_model,12
        jl    @f
        add    edi,[esp]
       @@:
 ,7 → 2931,7
        mov   byte[edi+2],0xff   ;al
        mov   word[eax],dx
        add   eax,2
-       cmp    [dr_flag],12
+       cmp   .dr_model,12
        jl    @f
        add   edi,4
        loop  .do
 ,7 → 3201,6
       .exit:
         mov     dword[edi],-1
 ret
 alloc_mem_for_tp:
         mov     eax, 68
         cmp     [re_alloc_flag],1
 ,7 → 3267,15
         mov     [points_rotated_ptr], eax
         mov     eax, 68
+        mov     ebx, 12
         mov     ecx, [points_count_var]
+        shl     ecx,2
+        mov     edx,[tex_points_ptr]
+        int     0x40
+        mov     [tex_points_ptr], eax
+        mov     eax, 68
+        mov     ecx, [points_count_var]
         inc     ecx
         shl     ecx, 3
         mov     edx,[points_translated_ptr]
 ,7 → 3284,6
 ret
 read_from_disk:
     mov     eax, 68
     mov     ebx, 11
 ,11 → 3329,11
         mov     edi,menu
       .again:
         mov     eax,8             ; function 8 : define and draw button
-        mov     bx,[size_x_var]
+        movzx   ebx,word[size_x_var]
         shl     ebx,16
         add     ebx,(10)*65536+62      ; [x start] *65536 + [x size]
         movzx   ecx,byte[edi]                 ; button id = position+2
-        sub     cl,2
+        sub     ecx,2
         lea     ecx,[ecx*5]
         lea     ecx,[ecx*3]
         add     ecx,25
 ,10 → 3345,10
          ; BUTTON  LABEL
         mov     eax,4                           ; function 4 : write text to window
         movzx   ebx,byte[edi]
-        sub     bl,2                            ; button id, according to position
+        sub     ebx,2                            ; button id, according to position
         lea     ebx,[ebx*3]
         lea     ebx,[ebx*5]
-        mov     cx,[size_x_var]
+        movzx   ecx,word[size_x_var]
         shl     ecx,16
         add     ebx,ecx
         add     ebx,(12)*65536+28        ; [x start] *65536 + [y start]
 ,6 → 3440,9
 ;   *******  WINDOW DEFINITIONS AND DRAW ********
 ;   *********************************************
     draw_window:
+        movzx   eax,[fire_flag]
+        push    eax
+    ;    int3
         mov     eax,12          ; function 12:tell os about windowdraw
         mov     ebx,1           ; 1, start of draw
         int     0x40
 ,6 → 3476,7
      ;   add     edx,130*65536+60  ; [x start] *65536 + [y start]
      ;   mov     esi,0x00ddeeff  ; font 1 & color ( 0xF0RRGGBB )
      ;   int     0x40
        call  write_info
         ; ADD VECTOR LABEL      ; add vector buttons - 30 ++
 ,6 → 3629,8
         mov     eax,12          ; function 12:tell os about windowdraw
         mov     ebx,2           ; 2, end of draw
         int     0x40
+        pop     eax
+        mov     [fire_flag],al
         ret

Subversion Repositories Kolibri OS

Compare Revisions

Ignore whitespace Rev 9236 → Rev 9237