WebSVN – Kolibri OS – Blame – /programs/demos/view3ds/3ray_shd.inc

Rev	Author	Line No.	Line
9237	leency	1	; Ray casted shadows
		2	; by Maciej Guba.
		3	; http://macgub.co.pl
		4
		5
		6	ROUND2 equ 10
		7	ray_shad:
		8	;--- Procedure render triangle with ray casted shadow ---
		9	;--- effect. Calc intersection with all triangles in ----
		10	;--- everypixel. Its not real time process, especially --
		11	;--- when many triangles are computed. ------------------
		12	;------in - eax - x1 shl 16 + y1 ------------------------
		13	;---------- ebx - x2 shl 16 + y2 ------------------------
		14	;---------- ecx - x3 shl 16 + y3 ------------------------
		15	;---------- edx - ptr to fur coords struct --------------
		16	;---------- esi - pointer to stencil / Z-buffer, filled -
		17	;-------------- with dword float variables, it masks --
		18	;-------------- 'Z' position (coord) of every front ---
		19	;-------------- pixel. --------------------------------
		20	;---------- edi - pointer to screen buffer --------------
		21	;---------- xmm0 - 1st normal vector --------------------
		22	;---------- xmm1 - 2cond normal vector ------------------
		23	;---------- xmm2 - 3rd normal vector --------------------
		24	;---------- xmm3 - --------------------------------------
		25	;---------- xmm4 - lo -> hi z1, z2, z3 coords -----------
		26	;--------------- as dwords floats ---------------------
		27	;---------- xmm5 - lo -> hi y_min, y_max, x_min, x_max --
		28	;--------------- as dword integers --------------------
		29	;-----------mm7 - current triangle index ---------------
		30	;---------------------- stack - no parameters -----------
		31	;--------------------------------------------------------
		32	;----------------- procedure don't save registers !! ----
		33
		34	push ebp
		35	mov ebp,esp
		36	sub esp,1024
		37	sub ebp,16
		38	and ebp,0xfffffff0
		39
		40	.1_nv equ [ebp-16]
		41	.2_nv equ [ebp-32]
		42	.3_nv equ [ebp-48]
		43	.l_v equ [ebp-64]
		44	.z3 equ [ebp-72]
		45	.z2 equ [ebp-76]
		46	.z1 equ [ebp-80]
		47	.x1 equ [ebp-82]
		48	.y1 equ [ebp-84]
		49	.x2 equ [ebp-86]
		50	.y2 equ [ebp-88]
		51	.x3 equ [ebp-90]
		52	.y3 equ [ebp-92]
		53	.Zbuf equ [ebp-96]
		54	.x_max equ [ebp-100]
		55	.x_min equ [ebp-104]
		56	.y_max equ [ebp-108]
		57	.y_min equ [ebp-112]
		58	.screen equ [ebp-116]
		59	.dx12 equ [ebp-120]
		60	.dx13 equ [ebp-124]
		61	.dx23 equ [ebp-128]
		62	.dn12 equ [ebp-144]
		63	.dn13 equ [ebp-160]
		64	.dn23 equ [ebp-176]
		65	.dz12 equ [ebp-180]
		66	.dz13 equ [ebp-184]
		67	.dz23 equ [ebp-188]
		68	.cnv1 equ [ebp-208] ; current normal vectors
		69	.cnv2 equ [ebp-240]
		70	.cz2 equ [ebp-244]
		71	.cz1 equ [ebp-248]
		72	.tri_no equ [ebp-252]
		73
		74
		75	.sort3: ; sort triangle coordinates...
		76	cmp ax,bx
		77	jle .sort1
		78	xchg eax,ebx
		79	shufps xmm4,xmm4,11100001b
		80	movaps xmm6,xmm0
		81	movaps xmm0,xmm1
		82	movaps xmm1,xmm6
		83
		84	.sort1:
		85	cmp bx,cx
		86	jle .sort2
		87	xchg ebx,ecx
		88	shufps xmm4,xmm4,11011000b
		89	movaps xmm6,xmm1
		90	movaps xmm1,xmm2
		91	movaps xmm2,xmm6
		92
		93	jmp .sort3
		94
		95	.sort2:
		96
		97	movaps .z1,xmm4
		98	mov .y1,eax
		99	mov .y2,ebx
		100	mov .y3,ecx
		101
		102	movdqa .y_min,xmm5
		103	if 1 ; check if at last only fragment
		104	packssdw xmm5,xmm5 ; of triangle is in visable area
		105	pshuflw xmm5,xmm5,11011000b
		106	movdqu xmm7,.y3
		107	movdqa xmm6,xmm5
		108	pshufd xmm5,xmm5,0 ; xmm5 lo-hi -> broadcasted y_min, x_min
		109	pshufd xmm6,xmm6,01010101b ;xmm6 -> brd y_max x_max
		110	movdqa xmm4,xmm7
		111	pcmpgtw xmm7,xmm5
		112	pcmpgtw xmm4,xmm6
		113	pxor xmm7,xmm4
		114	pmovmskb eax,xmm7
		115	and eax,0x00aaaaaa
		116	or eax,eax
		117	jz .rpt_loop2_end
		118	end if
		119	movd .tri_no,mm7
		120	movaps .1_nv,xmm0
		121	movaps .2_nv,xmm1
		122	movaps .3_nv,xmm2
		123	; movaps .l_v,xmm3
		124	mov .Zbuf,esi
		125	mov .screen,edi
		126
		127
		128
		129	mov bx,.y2 ; calc deltas
		130	sub bx,.y1
		131	jnz .rpt_dx12_make
		132
		133	xorps xmm7,xmm7
		134	mov dword .dx12,0
		135	mov dword .dz12,0
		136	movaps .dn12,xmm7
		137	jmp .rpt_dx12_done
		138
		139	.rpt_dx12_make:
		140	mov ax,.x2
		141	sub ax,.x1
		142	cwde
		143	movsx ebx,bx
		144	shl eax,ROUND2
		145	cdq
		146	idiv ebx
		147	mov .dx12,eax
		148
		149	cvtsi2ss xmm6,ebx
		150	movss xmm5,.z2
		151	rcpss xmm6,xmm6
		152	subss xmm5,.z1
		153	mulss xmm5,xmm6
		154	movss .dz12,xmm5
		155
		156	shufps xmm6,xmm6,0
		157	movaps xmm0,.2_nv
		158	subps xmm0,.1_nv
		159	mulps xmm0,xmm6
		160	movaps .dn12,xmm0
		161	; subps xmm3,xmm0
		162	; mulps xmm3,xmm6
		163
		164	.rpt_dx12_done:
		165	mov bx,.y3 ; calc deltas
		166	sub bx,.y1
		167	jnz .rpt_dx13_make
		168
		169	xorps xmm7,xmm7
		170	mov dword .dx13,0
		171	mov dword .dz13,0
		172	movaps .dn13,xmm7
		173	jmp .rpt_dx13_done
		174
		175	.rpt_dx13_make:
		176	mov ax,.x3
		177	sub ax,.x1
		178	cwde
		179	movsx ebx,bx
		180	shl eax,ROUND2
		181	cdq
		182	idiv ebx
		183	mov .dx13,eax
		184
		185	cvtsi2ss xmm6,ebx
		186	movss xmm5,.z3
		187	rcpss xmm6,xmm6
		188	subss xmm5,.z1
		189	mulss xmm5,xmm6
		190	movss .dz13,xmm5
		191
		192	movaps xmm0,.3_nv
		193	subps xmm0,.1_nv
		194	shufps xmm6,xmm6,0
		195	mulps xmm0,xmm6
		196	movaps .dn13,xmm0
		197
		198	; mulps xmm0,xmm6
		199
		200	.rpt_dx13_done:
		201
		202	mov bx,.y3 ; calc deltas
		203	sub bx,.y2
		204	jnz .rpt_dx23_make
		205
		206	xorps xmm7,xmm7
		207	mov dword .dx23,0
		208	mov dword .dz23,0
		209	movaps .dn23,xmm7
		210
		211	jmp .rpt_dx23_done
		212
		213	.rpt_dx23_make:
		214	mov ax,.x3
		215	sub ax,.x2
		216	cwde
		217	movsx ebx,bx
		218	shl eax,ROUND2
		219	cdq
		220	idiv ebx
		221	mov .dx23,eax
		222
		223	cvtsi2ss xmm6,ebx
		224	movss xmm5,.z3
		225	rcpss xmm6,xmm6
		226	subss xmm5,.z2
		227	mulss xmm5,xmm6
		228	movss .dz23,xmm5
		229
		230	movaps xmm0,.3_nv
		231	subps xmm0,.2_nv
		232	shufps xmm6,xmm6,0
		233	mulps xmm0,xmm6
		234	movaps .dn23,xmm0
		235	; mulps xmm0,xmm6
		236
		237	.rpt_dx23_done:
		238
		239	movsx eax,word .x1
		240	shl eax,ROUND2
		241	mov ebx,eax
		242	mov ecx,.z1
		243	mov .cz1,ecx
		244	mov .cz2,ecx
		245	movaps xmm0,.1_nv
		246	movaps .cnv1,xmm0
		247	movaps .cnv2,xmm0
		248	mov edi,.screen
		249	mov esi,.Zbuf
		250	movsx ecx,word .y1
		251	cmp cx,.y2
		252
		253	jge .rpt_loop1_end
		254
		255	.rpt_loop1:
		256	pushad
		257
		258	movaps xmm2,.y_min
		259	movaps xmm0,.cnv1
		260	movaps xmm1,.cnv2
		261	movlps xmm3,.cz1
		262	; movaps xmm4,.l_v
		263	sar ebx,ROUND2
		264	sar eax,ROUND2
		265	movd mm7,.tri_no
		266
		267	call ray_shd_l
		268
		269	popad
		270	movaps xmm0,.cnv1
		271	movaps xmm1,.cnv2
		272	; fur x,y
		273	movss xmm2,.cz1
		274	movss xmm3,.cz2
		275	shufps xmm4,xmm4,01001110b
		276	addps xmm0,.dn13
		277	addps xmm1,.dn12
		278	addss xmm2,.dz13
		279	addss xmm3,.dz12
		280
		281
		282	add eax,.dx13
		283	add ebx,.dx12
		284
		285	shufps xmm4,xmm4,01001110b
		286	movaps .cnv1,xmm0
		287	movaps .cnv2,xmm1
		288	movss .cz1,xmm2
		289	movss .cz2,xmm3
		290
		291	add ecx,1
		292	cmp cx,.y2
		293	jl .rpt_loop1
		294
		295
		296	.rpt_loop1_end:
		297	movsx ecx,word .y2
		298	cmp cx,.y3
		299	jge .rpt_loop2_end
		300
		301	movsx ebx,word .x2 ; eax - cur x1
		302	shl ebx,ROUND2 ; ebx - cur x2
		303	push dword .z2
		304	pop dword .cz2
		305	movaps xmm0,.2_nv
		306	movaps .cnv2,xmm0
		307
		308	mov edi,.screen
		309	mov esi,.Zbuf
		310
		311
		312	.rpt_loop2:
		313	pushad
		314	movaps xmm2,.y_min
		315	movaps xmm0,.cnv1
		316	movaps xmm1,.cnv2
		317	movlps xmm3,.cz1
		318	; movaps xmm4,.l_v
		319	sar ebx,ROUND2
		320	sar eax,ROUND2
		321	movd mm7,.tri_no
		322
		323	call ray_shd_l
		324
		325	popad
		326	movaps xmm0,.cnv1
		327	movaps xmm1,.cnv2
		328	movss xmm2,.cz1
		329	movss xmm3,.cz2
		330
		331	addps xmm0,.dn13
		332	addps xmm1,.dn23
		333	addss xmm2,.dz13
		334	addss xmm3,.dz23
		335	add eax,.dx13
		336	add ebx,.dx23
		337	addps xmm4,xmm6
		338
		339	movaps .cnv1,xmm0
		340	movaps .cnv2,xmm1
		341	movss .cz1,xmm2
		342	movss .cz2,xmm3
		343
		344	add ecx,1
		345	cmp cx,.y3
		346	jl .rpt_loop2
		347
		348	.rpt_loop2_end:
		349
		350	add esp,1024
		351	pop ebp
		352
		353
		354
		355	ret
9512	IgorA	356
9237	leency	357	ray_shd_l:
		358	; in:
		359	; xmm0 - normal vector 1
		360	; xmm1 - normal vect 2
		361	; xmm3 - lo -> hi z1, z2 coords as dwords floats
		362	; xmm2 - lo -> hi y_min, y_max, x_min, x_max
		363	; as dword integers
		364	; xmm4 - ----
		365	; mm7 - current triangle index
		366	; eax - x1
		367	; ebx - x2
		368	; ecx - y
		369	; edx - -----
		370	; edi - screen buffer
		371	; esi - z buffer / stencil buffer filled with dd floats
		372
		373	push ebp
		374	mov ebp,esp
9740	macgub	375	sub esp,270
9237	leency	376	sub ebp,16
		377	and ebp,0xfffffff0
		378
		379	.n1 equ [ebp-16]
		380	.n2 equ [ebp-32]
		381	.lv equ [ebp-48]
		382	.lx1 equ [ebp-52]
		383	.lx2 equ [ebp-56]
		384	.z2 equ [ebp-60]
		385	.z1 equ [ebp-64]
		386	.screen equ [ebp-68]
		387	.zbuff equ [ebp-72]
		388	.x_max equ [ebp-74]
		389	.x_min equ [ebp-76]
		390	.y_max equ [ebp-78]
		391	.y_min equ [ebp-80]
		392	.dn equ [ebp-96]
		393	.dz equ [ebp-100]
		394	.y equ [ebp-104]
9512	IgorA	395	.startx equ [ebp-108]
9237	leency	396	.cnv equ [ebp-128]
		397	.Rlen equ [ebp-128-16]
		398	.r1 equ [ebp-128-32]
		399	.vect_t equ [ebp-128-48]
		400	.cur_tri equ [ebp-128-64]
		401	; .p3t equ [ebp-128-80]
		402	.nray equ [ebp-128-96]
		403	.final_col equ [ebp-128-112]
		404	.aabb_mask equ dword[ebp-128-112-4]
		405
		406	mov .y,ecx
		407	movdqa xmm4,xmm2
		408	packssdw xmm2,xmm2
		409	movq .y_min,xmm2
		410	cmp cx,.y_min
		411	jl .end_rp_line
		412	cmp cx,.y_max
		413	jge .end_rp_line ;
		414	cmp eax,ebx
		415	je .end_rp_line
		416	jl @f
		417	xchg eax,ebx
		418	movaps xmm7,xmm0
		419	movaps xmm0,xmm1
		420	movaps xmm1,xmm7
		421	shufps xmm3,xmm3,11100001b
		422	@@:
		423	movd .cur_tri,mm7
9740	macgub	424	; sub .cur_tri,dword 1
9237	leency	425	cmp ax,.x_max
		426	jge .end_rp_line
		427	cmp bx,.x_min
		428	jle .end_rp_line
		429	; movaps .lv,xmm4
		430	andps xmm0,[zero_hgst_dd]
		431	andps xmm1,[zero_hgst_dd]
		432	movaps .n1,xmm0
		433	movaps .n2,xmm1
		434	mov .lx1,eax
9512	IgorA	435	; mov .startx,eax
9237	leency	436	mov .lx2,ebx
		437	movlps .z1,xmm3
		438
		439	sub ebx,eax
		440	cvtsi2ss xmm7,ebx
		441	rcpss xmm7,xmm7
		442	shufps xmm7,xmm7,0
		443	subps xmm1,xmm0
		444	mulps xmm1,xmm7
		445	movaps .dn,xmm1
		446	shufps xmm3,xmm3,11111001b
		447	subss xmm3,.z1
		448	mulss xmm3,xmm7
		449	movss .dz,xmm3
		450
		451	subps xmm6,xmm5
		452	mulps xmm6,xmm7
		453
		454	mov ebx,.lx1
		455	cmp bx,.x_min ; clipping on function4
		456	jge @f
		457	movzx eax,word .x_min
		458	sub eax,ebx
		459	cvtsi2ss xmm7,eax
		460	shufps xmm7,xmm7,0
		461	mulss xmm3,xmm7
		462	mulps xmm1,xmm7
		463	mulps xmm6,xmm7
		464	addss xmm3,.z1
		465	addps xmm1,.n1
		466	addps xmm6,xmm5
		467	movsx eax,word .x_min
		468	movss .z1,xmm3
		469	movaps .n1,xmm1
		470	mov dword .lx1,eax
		471	@@:
		472
		473	movzx eax,word .x_max
		474	cmp .lx2,eax
		475	jl @f
		476	mov .lx2,eax
		477	@@:
		478	movzx eax,word[xres_var]
		479	mul dword .y
		480	add eax,.lx1
		481	mov .zbuff,esi
		482	mov .screen,edi
		483	shl eax,2
		484	add edi,eax
		485	add esi,eax
		486	mov ecx,.lx2
		487	sub ecx,.lx1
		488
		489	movd xmm0,[vect_x]
		490	punpcklwd xmm0,[the_zero]
		491	cvtdq2ps xmm0,xmm0
		492	movaps .vect_t,xmm0
		493
		494
		495	.ddraw:
		496
		497	xorps xmm0,xmm0
		498	movss xmm2,.z1
		499	movss xmm5,.z1
		500	movaps .final_col,xmm0
		501	addss xmm2,[f1]
		502	subss xmm5,[f1]
		503	cmpnltss xmm2,dword[esi]
		504	cmpnltss xmm5,dword[esi]
		505	pxor xmm2,xmm5
		506	movd eax,xmm2
		507	or eax,eax
		508	jz .skips
		509
		510	movaps xmm7,.n1
		511	andps xmm7,[zero_hgst_dd]
		512	mulps xmm7,xmm7 ; normalize
		513	haddps xmm7,xmm7
		514	haddps xmm7,xmm7
		515	rsqrtps xmm7,xmm7
		516	mulps xmm7,.n1
		517	movaps .cnv,xmm7
		518	mov ebx,point_light_coords
		519	mov edx,lights_aligned
		520	xor eax,eax
		521	.nx_light:
		522	pushad
		523	cvtsi2ss xmm0,.lx1
		524	cvtsi2ss xmm1,.y
		525	movss xmm2,.z1
		526	movlhps xmm0,xmm1
		527	shufps xmm0,xmm2,11001000b
		528	subps xmm0,[ebx] ; xmm0 - ray end, -> current vertex
		529	movaps xmm3,[ebx]
		530	andps xmm0,[zero_hgst_dd]
		531	movaps xmm1,xmm0
		532	mulps xmm0,xmm0
		533	haddps xmm0,xmm0
		534	haddps xmm0,xmm0
		535	sqrtps xmm0,xmm0
		536	movss .Rlen,xmm0
		537	rcpps xmm0,xmm0
		538	mulps xmm0,xmm1 ; xmm0 - normalized ray vector
		539	andps xmm0,[zero_hgst_dd]
		540	movaps .nray,xmm0
		541	movaps .r1,xmm3 ; ray orgin
		542	if 0
		543	movaps xmm1,xmm3
		544	call calc_bounding_box
		545
		546	mov .aabb_mask,eax
		547	end if
		548	mov edi,[triangles_ptr]
		549	xor ecx,ecx
		550	.nx_tri: ; next triangle
9512	IgorA	551	; mov eax,.lx1
		552	; cmp eax,.startx
		553	; je @f ; prevent artifact borders on tri
		554	; cmp eax,.lx2 ; NOT work as I want !!
		555	; je @f
9237	leency	556
		557	cmp ecx,.cur_tri ; prevent self shadowing
		558	je .skipp
9512	IgorA	559	@@:
9237	leency	560	if 0
		561	mov edi,ecx
		562	imul edi,[i12]
		563	add edi,[triangles_ptr]
		564	mov eax,[edi]
		565	mov ebx,[edi+4]
		566	mov edx,[edi+8]
		567	imul eax,[i12]
		568	imul ebx,[i12]
		569	imul edx,[i12]
		570	add eax,[points_ptr]
		571	add ebx,[points_ptr]
		572	add edx,[points_ptr]
		573	movups xmm2,[eax]
		574	movups xmm3,[ebx]
		575	movups xmm4,[edx]
		576	andps xmm2,[sign_mask]
		577	andps xmm3,[sign_mask]
		578	andps xmm4,[sign_mask]
		579	movmskps ebx,xmm4
		580	cmpeqps xmm2,xmm3
		581	cmpeqps xmm3,xmm4
		582	andps xmm2,xmm3
		583	movmskps eax,xmm2
		584	and eax,111b
		585	and ebx,111b
		586	cmp eax,111b
		587	jne @f
		588	bt .aabb_mask,ebx
		589	jnc .skipp
		590	@@:
		591	end if
		592	mov edi,ecx
		593	imul edi,[i12]
		594	add edi,[triangles_ptr]
		595	mov eax,[edi]
		596	mov ebx,[edi+4]
		597	mov edx,[edi+8]
		598	imul eax,[i12]
		599	imul ebx,[i12]
		600	imul edx,[i12]
		601	add eax,[points_rotated_ptr]
		602	add ebx,[points_rotated_ptr]
		603	add edx,[points_rotated_ptr]
		604	movups xmm2,[eax]
		605	movups xmm3,[ebx]
		606	movups xmm4,[edx]
		607	addps xmm2,.vect_t
		608	addps xmm3,.vect_t
		609	addps xmm4,.vect_t
		610
		611
		612	;intersect_tri: procs header
		613	; in:
		614	; xmm0 - ray direction ; should be normalized
		615	; xmm1 - ray orgin
		616	; xmm2 - tri vert1
		617	; xmm3 - tri vert2
		618	; xmm4 - tri vert3
		619	; if eax = 1 - intersction with edge
		620	; xmm6 - edge lenght
		621	; if eax = 0 - intersect with ray (classic)
		622	; out:
		623	; eax = 1 - intersection occured
		624	; xmm0 - float lo -> hi = t, v, u, ...
		625
		626	movss xmm6,.Rlen
		627	movaps xmm0,.nray
		628	movaps xmm1,.r1
		629	subss xmm6,[the_one]
		630	mov eax,1
		631	push ecx
		632	call intersect_tri
		633	pop ecx
		634	cmp eax,1
		635	je .inter
		636	.skipp:
		637	.skp:
		638	inc ecx
		639	cmp ecx,[triangles_count_var]
		640	jnz .nx_tri
		641	; jz .do_process
		642	; comiss xmm0,.Rlen
		643	; jl .inter
		644
		645	popad
		646	.do_process:
		647	movaps xmm5,.nray ;[edx]
		648	andps xmm5,[zero_hgst_dd] ; global
		649	mulps xmm5,.cnv ;.lv ; last dword should be zeroed
		650	; andps xmm5,[sign_z] ; global
		651	haddps xmm5,xmm5
		652	haddps xmm5,xmm5
		653	andps xmm5,[abs_mask] ; global
		654	movaps xmm7,xmm5
		655	mulps xmm7,xmm7
		656	mulps xmm7,xmm7
		657	mulps xmm5,[edx+16]
		658	mulps xmm7,xmm7
		659	mulps xmm7,xmm7
		660	mulps xmm7,[edx+48]
		661	addps xmm5,xmm7
		662	minps xmm5,[mask_255f] ; global
		663	maxps xmm5,.final_col ; addps maxps
		664	movaps .final_col,xmm5
		665	jmp .nx_loop
		666	.inter:
		667
		668	popad
		669	.nx_loop:
		670	; add edx,64 ; unncomment to achive 3 lights
		671	; add ebx,16
		672	; cmp edx,lights_aligned_end ; global
		673	; jnz .nx_light
		674
		675	movaps xmm1,.final_col
		676	cvtps2dq xmm1,xmm1
		677	packssdw xmm1,xmm1
		678	packuswb xmm1,xmm1
		679	movd [edi],xmm1
		680	.skips:
		681	movaps xmm0,.n1
		682	movss xmm2,.z1
		683	add edi,4
		684	add esi,4
		685	add dword .lx1,1
		686	addps xmm0,.dn
		687	addss xmm2,.dz
		688	movaps .n1,xmm0
		689	movss .z1,xmm2
		690	dec ecx
		691	jnz .ddraw
		692	.end_rp_line:
9740	macgub	693	add esp,270
9237	leency	694	pop ebp
		695
		696	ret

Subversion Repositories Kolibri OS

(root)/programs/demos/view3ds/3ray_shd.inc – Rev 9740