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

Rev	Author	Line No.	Line
8719	leency	1	x3d equ 0
		2	y3d equ 2
		3	z3d equ 4
		4	vec_x equ 0
		5	vec_y equ 4
		6	vec_z equ 8
9237	leency	7
		8	if 0 ; Ext >= SSE3
		9	calc_bounding_box:
		10	; in:
		11	; xmm0 - normal vector of ray
		12	; xmm1 - light origin
		13	; out:
		14	; eax - axis aligned bounding boxes bit mask
		15
		16	.rmx equ [ebp-36]
		17	.nray equ [ebp-64]
		18	.origin equ [ebp-80]
		19	.dirfrac equ [ebp-96]
		20	.nrayr equ [ebp-112]
		21	.originr equ [ebp-128]
		22	.tmin equ [ebp-132]
		23	.tmax equ [ebp-136]
		24
		25
		26	push ebp
		27	mov ebp,esp
		28	and ebp,-16
		29	sub esp,160
		30
		31	movss xmm5,[rsscale]
		32	shufps xmm5,xmm1,0
		33	movd xmm2,[vect_x]
		34	punpcklwd xmm2,[the_zero]
		35	cvtdq2ps xmm2,xmm2
		36	subps xmm1,xmm2
		37	movaps .origin,xmm1
		38	mulps xmm0,xmm5
		39	movaps .nray,xmm0
		40
		41	mov esi,matrix
		42	lea edi,.rmx
		43	call reverse_mx_3x3
		44
		45	; in: esi - ptr to points(normals], each point(normal) coeficient as dword
		46	; edi - ptr to rotated points(normals)
		47	; ebx - ptr to 3x3 (9 dwords, 36 bytes) rotation matrix
		48	; ecx - number of points(normals)
		49
		50	; reverse transform
		51	lea esi,.nray
		52	lea edi,.nrayr
		53	lea ebx,.rmx
		54	mov ecx,1
		55	call rotary
		56
		57	lea esi,.origin
		58	lea edi,.originr
		59	lea ebx,.rmx
		60	mov ecx,1
		61	call rotary
		62
		63	xor ecx,ecx
		64	mov ebx,aabb1
		65	xor eax,eax
		66	rcpps xmm7,.nrayr
		67	movaps .dirfrac,xmm7
		68
		69	.nx_aabb:
		70	movaps xmm5,[ebx]
		71	movaps xmm6,[ebx]
		72	minps xmm5,[the_zero]
		73	maxps xmm6,[the_zero]
		74	; xmm5 - lb corner of AABB with minimal coordinates
		75	; xmm6 - rt cor. of AABB wit maximum coords
		76	subps xmm5,.originr
		77	subps xmm6,.originr
		78	mulps xmm5,.dirfrac ; xmm5 - tx1, ty1
		79	mulps xmm6,.dirfrac ; xmm6 - tx2, ty2
		80	movaps xmm1,xmm6
		81	movaps xmm2,xmm6
		82
		83
		84	minps xmm1,xmm5
		85	maxps xmm2,xmm5
		86
		87	movaps xmm5,xmm1
		88	movaps xmm6,xmm2
		89	shufps xmm5,xmm5,11100001b
		90	shufps xmm6,xmm6,11100001b
		91	maxss xmm1,xmm5 ;t min
		92	minss xmm2,xmm6 ;t max
		93	comiss xmm2,xmm1
		94	jb .no_inter
		95	.yes:
		96	bts eax,ecx
		97	.no_inter:
		98	add ebx,16
		99	inc ecx
		100	cmp ecx,8
		101	jne .nx_aabb
		102
		103	; out: eax - bit mask
		104	add esp,160
		105	pop ebp
		106	ret
		107	end if
		108
8719	leency	109	reverse_mx_3x3:
		110	; esi - source matrix
		111	; edi - desired reversed matrix
		112
		113	push ebp
		114	mov ebp,esp
		115	sub esp,4
		116	.det equ ebp-4
		117
		118	fninit
		119	fld dword[esi]
		120	fmul dword[esi+16]
		121	fmul dword[esi+32]
		122	fld dword[esi+12]
		123	fmul dword[esi+28]
		124	fmul dword[esi+8]
		125	faddp
		126	fld dword[esi+24]
		127	fmul dword[esi+4]
		128	fmul dword[esi+20]
		129	faddp
		130	fld dword[esi]
		131	fmul dword[esi+28]
		132	fmul dword[esi+20]
		133	fchs
		134	faddp
		135	fld dword[esi+24]
		136	fmul dword[esi+16]
		137	fmul dword[esi+8]
		138	fchs
		139	faddp
		140	fld dword[esi+12]
		141	fmul dword[esi+4]
		142	fmul dword[esi+32]
		143	fchs
		144	faddp
		145	fstp dword[.det]
		146	cmp dword[.det],0
		147	jne @f
		148	int3
		149	@@:
		150	; fld1
		151	; fdiv dword[.det]
		152	; fstp dword[.det]
		153
		154	fld dword[esi+16]
		155	fmul dword[esi+32]
		156	fld dword[esi+20]
		157	fmul dword[esi+28]
		158	fchs
		159	faddp
		160	fdiv dword[.det]
		161	fstp dword[edi]
		162
		163	fld dword[esi+8]
		164	fmul dword[esi+28]
		165	fld dword[esi+4]
		166	fmul dword[esi+32]
		167	fchs
		168	faddp
		169	fdiv dword[.det]
		170	fstp dword[edi+4]
		171
		172	fld dword[esi+4]
		173	fmul dword[esi+20]
		174	fld dword[esi+8]
		175	fmul dword[esi+16]
		176	fchs
		177	faddp
		178	fdiv dword[.det]
		179	fstp dword[edi+8]
		180
		181	fld dword[esi+20]
		182	fmul dword[esi+24]
		183	fld dword[esi+12]
		184	fmul dword[esi+32]
		185	fchs
		186	faddp
		187	fdiv dword[.det]
		188	fstp dword[edi+12]
		189
		190	fld dword[esi]
		191	fmul dword[esi+32]
		192	fld dword[esi+8]
		193	fmul dword[esi+24]
		194	fchs
		195	faddp
		196	fdiv dword[.det]
		197	fstp dword[edi+16]
		198
		199	fld dword[esi+8]
		200	fmul dword[esi+12]
		201	fld dword[esi]
		202	fmul dword[esi+20]
		203	fchs
		204	faddp
		205	fdiv dword[.det]
		206	fstp dword[edi+20]
		207
		208	fld dword[esi+12]
		209	fmul dword[esi+28]
		210	fld dword[esi+16]
		211	fmul dword[esi+24]
		212	fchs
		213	faddp
		214	fdiv dword[.det]
		215	fstp dword[edi+24]
		216
		217	fld dword[esi+4]
		218	fmul dword[esi+24]
		219	fld dword[esi]
		220	fmul dword[esi+28]
		221	fchs
		222	faddp
		223	fdiv dword[.det]
		224	fstp dword[edi+28]
		225
		226	fld dword[esi]
		227	fmul dword[esi+16]
		228	fld dword[esi+4]
		229	fmul dword[esi+12]
		230	fchs
		231	faddp
		232	fdiv dword[.det]
		233	fstp dword[edi+32]
		234
		235
		236	mov esp,ebp
		237	pop ebp
		238	ret
9237	leency	239	; 3d point - triple integer word coordinate
		240	; vector - triple float dword coordinate
		241	;----------------------in: --------------------------------
		242	;------------------------ esi - pointer to 1st 3d point ---
		243	;------------------------ edi - pointer to 2nd 3d point ---
		244	;------------------------ ebx - pointer to result vector --
		245	;---------------------- out : none ------------------------
8719	leency	246
		247	make_vector_r:
		248	if Ext < SSE2
		249	fninit
		250	fld dword[edi] ;edi+x3d
		251	fsub dword[esi] ;esi+x3d
		252	fstp dword[ebx+vec_x]
		253
		254	fld dword[edi+4]
		255	fsub dword[esi+4]
		256	fstp dword[ebx+vec_y]
		257
		258	fld dword[edi+8]
		259	fsub dword[esi+8]
		260	fstp dword[ebx+vec_z]
		261	else
		262	movups xmm0,[esi]
		263	movups xmm1,[edi]
		264	subps xmm1,xmm0
		265	movlps [ebx],xmm1
		266	movhlps xmm1,xmm1
		267	movss [ebx+8],xmm1
		268	end if
		269
		270	ret
		271	;---------------------- in: -------------------------------
		272	;--------------------------- esi - pointer to 1st vector --
		273	;--------------------------- edi - pointer to 2nd vector --
		274	;--------------------------- ebx - pointer to result vector
		275	;---------------------- out : none
		276	cross_product:
		277	fninit
		278	fld dword [esi+vec_y]
		279	fmul dword [edi+vec_z]
		280	fld dword [esi+vec_z]
		281	fmul dword [edi+vec_y]
		282	fsubp ;st1 ,st
		283	fstp dword [ebx+vec_x]
		284
		285	fld dword [esi+vec_z]
		286	fmul dword [edi+vec_x]
		287	fld dword [esi+vec_x]
		288	fmul dword [edi+vec_z]
		289	fsubp ;st1 ,st
		290	fstp dword [ebx+vec_y]
		291
		292	fld dword [esi+vec_x]
		293	fmul dword [edi+vec_y]
		294	fld dword [esi+vec_y]
		295	fmul dword [edi+vec_x]
		296	fsubp ;st1 ,st
		297	fstp dword [ebx+vec_z]
		298	ret
9237	leency	299	cross_aligned:
		300	movaps xmm0,[esi]
		301	movaps xmm1,[esi]
		302	movaps xmm2,[edi]
		303	movaps xmm3,[edi]
		304	shufps xmm0,xmm0,00001001b
		305	shufps xmm1,xmm1,00010010b
		306	shufps xmm2,xmm2,00010010b
		307	shufps xmm3,xmm3,00001001b
		308	mulps xmm0,xmm2
		309	mulps xmm1,xmm3
		310	subps xmm0,xmm1
		311	movaps [ebx],xmm0
		312	ret
8719	leency	313	;----------------------- in: ------------------------------
		314	;---------------------------- edi - pointer to vector -----
		315	;----------------------- out : none
		316	normalize_vector:
9237	leency	317	if Ext >= SSE2
8719	leency	318	movups xmm0,[edi]
		319	andps xmm0,[zero_hgst_dd]
		320	movups xmm1,xmm0
		321	mulps xmm0,xmm0
9237	leency	322	movhlps xmm2,xmm0
		323	addps xmm0,xmm2
		324	movaps xmm2,xmm0
		325	shufps xmm2,xmm2,11100101b
		326	addps xmm0,xmm2
		327	shufps xmm0,xmm0,0
		328	; haddps xmm0,xmm0
		329	; haddps xmm0,xmm0
8719	leency	330	rsqrtps xmm0,xmm0
		331	mulps xmm0,xmm1
		332	movlps [edi],xmm0
		333	movhlps xmm0,xmm0
		334	movss [edi+8],xmm0
		335	else
		336
		337	fninit
		338	fld dword [edi+vec_x]
		339	fmul st, st
		340	fld dword [edi+vec_y]
		341	fmul st, st
		342	fld dword [edi+vec_z]
		343	fmul st, st
		344	faddp st1, st
		345	faddp st1, st
		346	fsqrt
		347
		348	ftst
		349	fstsw ax
		350	sahf
		351	jnz @f
		352
		353	fst dword [edi+vec_x]
		354	fst dword [edi+vec_y]
		355	fstp dword [edi+vec_z]
		356	ret
		357	@@:
		358	fld st
		359	fld st
		360	fdivr dword [edi+vec_x]
		361	fstp dword [edi+vec_x]
		362	fdivr dword [edi+vec_y]
		363	fstp dword [edi+vec_y]
		364	fdivr dword [edi+vec_z]
		365	fstp dword [edi+vec_z]
		366	end if
		367	ret
		368	;------------------in: -------------------------
		369	;------------------ esi - pointer to 1st vector
		370	;------------------ edi - pointer to 2nd vector
		371	;------------------out: ------------------------
		372	;------------------ st0 - dot-product
		373	dot_product:
		374	fninit
		375	;if Ext >=SSE3
		376	; movups xmm0,[esi]
		377	; movups xmm1,[edi]
		378	; andps xmm0,[zero_hgst_dd]
		379	; mulps xmm0,xmm1
		380	; haddps xmm0,xmm0
		381	; haddps xmm0,xmm0
		382	; movss [esp-4],xmm0
		383	; fld dword[esp-4]
		384	;else
		385	fld dword [esi+vec_x]
		386	fmul dword [edi+vec_x]
		387	fld dword [esi+vec_y]
		388	fmul dword [edi+vec_y]
		389	fld dword [esi+vec_z]
		390	fmul dword [edi+vec_z]
		391	faddp
		392	faddp
		393	;end if
		394	ret
		395
		396	; DOS version Coded by Mikolaj Felix aka Majuma
		397	; mfelix@polbox.com
		398	; www.majuma.xt.pl
		399	; into FASM translation by Macgub
		400	init_sincos_tab:
		401	.counter equ dword [ebp-4] ; cur angle
		402
		403	push ebp
		404	mov ebp,esp
		405
		406	xor eax,eax
		407	push eax ; init .counter
		408	mov edi,cos_tab
		409	mov esi,sin_tab
		410	mov ecx,256
		411	fninit
		412
		413	fld .counter
		414	@@:
		415	fld st
		416	fsincos
		417	fstp dword [edi]
		418	fstp dword [esi]
		419	; fadd [piD180]
		420	fadd [piD128]
		421	add esi,4
		422	add edi,4
		423	loop @b
		424	ffree st
		425
		426	mov esp,ebp
		427	pop ebp
		428	ret
		429	;------
		430	; esi - offset (pointer) to angles, edi offset to 3x3 matrix
		431	make_rotation_matrix:
		432	.sinx equ dword[ebp-4]
		433	.cosx equ dword[ebp-8]
		434	.siny equ dword[ebp-12]
		435	.cosy equ dword[ebp-16]
		436	.sinz equ dword[ebp-20]
		437	.cosz equ dword[ebp-24]
		438	push ebp
		439	mov ebp,esp
		440	sub esp,24
		441
		442	movzx ebx,word[esi]
		443	shl ebx,2
		444	mov eax,dword[sin_tab+ebx]
		445	mov .sinx,eax
		446	mov edx,dword[cos_tab+ebx]
		447	mov .cosx,edx
		448
		449	movzx ebx,word[esi+2]
		450	shl ebx,2
		451	mov eax,dword[sin_tab+ebx]
		452	mov .siny,eax
		453	mov edx,dword[cos_tab+ebx]
		454	mov .cosy,edx
		455
		456	movzx ebx,word[esi+4]
		457	shl ebx,2
		458	mov eax,dword[sin_tab+ebx]
		459	mov .sinz,eax
		460	mov edx,dword[cos_tab+ebx]
		461	mov .cosz,edx
		462
		463	fninit
		464	fld .cosy
		465	fmul .cosz
		466	fstp dword[edi]
		467
		468	fld .sinx
		469	fmul .siny
		470	fmul .cosz
		471	fld .cosx
		472	fmul .sinz
		473	fchs
		474	faddp
		475	fstp dword[edi+12]
		476
		477	fld .cosx
		478	fmul .siny
		479	fmul .cosz
		480	fld .sinx
		481	fmul .sinz
		482	faddp
		483	fstp dword[edi+24]
		484
		485	fld .cosy
		486	fmul .sinz
		487	fstp dword[edi+4]
		488
		489	fld .sinx
		490	fmul .siny
		491	fmul .sinz
		492	fld .cosx
		493	fmul .cosz
		494	faddp
		495	fstp dword[edi+16]
		496
		497	fld .cosx
		498	fmul .siny
		499	fmul .sinz
		500	fld .sinx
		501	fchs
		502	fmul .cosz
		503	faddp
		504	fstp dword[edi+28]
		505
		506	fld .siny
		507	fchs
		508	fstp dword[edi+8]
		509
		510	fld .cosy
		511	fmul .sinx
		512	fstp dword[edi+20]
		513
		514	fld .cosx
		515	fmul .cosy
		516	fstp dword[edi+32]
		517
		518	mov esp,ebp
		519	pop ebp
		520	ret
		521	;---------------------
		522	; in: esi - ptr to points(normals], each point(normal) coeficient as dword
		523	; edi - ptr to rotated points(normals)
		524	; ebx - ptr to 3x3 (9 dwords, 36 bytes) rotation matrix
		525	; ecx - number of points(normals)
		526	rotary:
		527	if Ext
		528	fninit
		529	.again:
		530
		531	fld dword[esi]
		532	fmul dword[ebx]
		533	fld dword[esi+4]
		534	fmul dword[ebx+12]
		535	faddp
		536	fld dword[esi+8]
		537	fmul dword[ebx+24]
		538	faddp
		539	fstp dword[edi]
		540
		541
		542	fld dword[esi+4]
		543	fmul dword[ebx+16]
		544	fld dword[esi]
		545	fmul dword[ebx+4]
		546	faddp
		547	fld dword[esi+8]
		548	fmul dword[ebx+28]
		549	faddp
		550	fstp dword[edi+4]
		551
		552
		553	fld dword[esi+8]
		554	fmul dword[ebx+32]
		555	fld dword[esi]
		556	fmul dword[ebx+8]
		557	fld dword[esi+4]
		558	fmul dword[ebx+20]
		559	faddp
		560	faddp
		561	fstp dword[edi+8]
		562
		563
		564	add esi,12
		565	add edi,12
		566	loop .again
		567	mov [edi],dword -1
		568	else
		569	; Copyright (C) 1999-2001 Brian Paul
		570	; Copyright (C) Maciej Guba
		571	;---------------------
		572	; in: esi - ptr to points(normals], each point(normal) coeficient as dword
		573	; edi - ptr to rotated points(normals)
		574	; ebx - ptr to 3x3 (9 dwords, 36 bytes) rotation matrix
		575	; ecx - number of points(normals)
		576	;align 32
		577	movups xmm4,[ebx]
		578	; lddqu xmm4,[ebx] ; I tried sse3 :D
		579	movups xmm5,[ebx+12]
		580	movups xmm6,[ebx+24]
		581	;align 32
		582	.again:
		583	movss xmm0,dword[esi]
		584	shufps xmm0,xmm0,0
		585	mulps xmm0,xmm4
		586
		587	movss xmm1,dword[esi+4]
		588	shufps xmm1,xmm1,0
		589	mulps xmm1,xmm5
		590
		591	movss xmm2,dword[esi+8]
		592	shufps xmm2,xmm2,0
		593	mulps xmm2,xmm6
		594
		595	addps xmm0,xmm1
		596	addps xmm0,xmm2
		597
		598	movups [edi],xmm0
		599
		600	add esi,12
		601	add edi,12
		602	dec ecx
		603	jne .again
		604	mov [edi],dword -1
		605	end if
		606	ret
		607	;----------------------------------------------
		608	; esi - pointer to 3x3 matrix
		609	add_scale_to_matrix:
		610	fninit
		611	fld [rsscale]
		612	fld dword[esi] ;-----
		613	fmul st,st1
		614	fstp dword[esi]
		615	fld dword[esi+12] ; x scale
		616	fmul st,st1
		617	fstp dword[esi+12]
		618	fld dword[esi+24]
		619	fmul st,st1
		620	fstp dword[esi+24] ;------
		621
		622	fld dword[esi+4] ;-----
		623	fmul st,st1
		624	fstp dword[esi+4]
		625	fld dword[esi+16] ; y scale
		626	fmul st,st1
		627	fstp dword[esi+16]
		628	fld dword[esi+28]
		629	fmul st,st1
		630	fstp dword[esi+28] ;------
		631
		632
		633	fld dword[esi+8] ;-----
		634	fmul st,st1
		635	fstp dword[esi+8]
		636	fld dword[esi+20] ; z scale
		637	fmul st,st1
		638	fstp dword[esi+20]
		639	fld dword[esi+32]
		640	fmulp st1,st
		641	fstp dword[esi+32] ;------
		642
		643	ret
		644
		645	;in esi - offset to 3d points (point as 3 dwords float)
		646	; edi - offset to 2d points ( as 3 words integer)
		647	; ecx - number of points
		648	translate_points: ; just convert into integer; z coord still needed
		649	fninit
		650	.again:
		651	if 0
		652	fld dword[esi+8]
		653	; fmul [rsscale]
		654	fist word[edi+4]
		655
		656	fisub [zobs]
		657	fchs
		658
		659	fld dword[esi]
		660	; fmul [rsscale]
		661	fisub [xobs]
		662	fimul [zobs]
		663	fdiv st0,st1
		664
		665	fiadd [xobs]
		666	fiadd [vect_x]
		667	fistp word[edi]
		668
		669	fld dword[esi+4]
		670	; fmul [rsscale]
		671	fisub [yobs]
		672	fimul [zobs]
		673	fdivrp ; st0,st1
		674
		675	fiadd [yobs]
		676	fiadd [vect_y]
		677	fistp word[edi+2]
		678	end if
		679	; movups xmm0,[esi]
		680	; cvtps2dq xmm0,xmm0
		681	; packsdw xmm0,xmm0
		682	; movq [edi]
		683	fld dword[esi]
9237	leency	684	fiadd word[vect_x]
8719	leency	685	fistp word[edi]
		686	fld dword[esi+4]
		687	fiadd [vect_y]
		688	fistp word[edi+2]
		689	fld dword[esi+8]
		690	fistp word[edi+4]
		691	add esi,12
		692	add edi,6
		693	dec ecx
		694	jnz .again
		695
		696	ret

Subversion Repositories Kolibri OS

(root)/programs/demos/view3ds/3dmath.inc – Rev 9237