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:
9512	IgorA	248	if Ext < SSE
8719	leency	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:
9512	IgorA	300	; params as above cross_p
9237	leency	301	movaps xmm0,[esi]
		302	movaps xmm1,[esi]
		303	movaps xmm2,[edi]
		304	movaps xmm3,[edi]
		305	shufps xmm0,xmm0,00001001b
		306	shufps xmm1,xmm1,00010010b
		307	shufps xmm2,xmm2,00010010b
		308	shufps xmm3,xmm3,00001001b
		309	mulps xmm0,xmm2
		310	mulps xmm1,xmm3
		311	subps xmm0,xmm1
		312	movaps [ebx],xmm0
		313	ret
8719	leency	314	;----------------------- in: ------------------------------
		315	;---------------------------- edi - pointer to vector -----
		316	;----------------------- out : none
		317	normalize_vector:
9237	leency	318	if Ext >= SSE2
8719	leency	319	movups xmm0,[edi]
		320	andps xmm0,[zero_hgst_dd]
		321	movups xmm1,xmm0
		322	mulps xmm0,xmm0
9237	leency	323	movhlps xmm2,xmm0
		324	addps xmm0,xmm2
		325	movaps xmm2,xmm0
		326	shufps xmm2,xmm2,11100101b
		327	addps xmm0,xmm2
		328	shufps xmm0,xmm0,0
		329	; haddps xmm0,xmm0
		330	; haddps xmm0,xmm0
8719	leency	331	rsqrtps xmm0,xmm0
		332	mulps xmm0,xmm1
		333	movlps [edi],xmm0
		334	movhlps xmm0,xmm0
		335	movss [edi+8],xmm0
		336	else
		337
		338	fninit
		339	fld dword [edi+vec_x]
		340	fmul st, st
		341	fld dword [edi+vec_y]
		342	fmul st, st
		343	fld dword [edi+vec_z]
		344	fmul st, st
		345	faddp st1, st
		346	faddp st1, st
		347	fsqrt
		348
		349	ftst
		350	fstsw ax
		351	sahf
		352	jnz @f
		353
		354	fst dword [edi+vec_x]
		355	fst dword [edi+vec_y]
		356	fstp dword [edi+vec_z]
		357	ret
		358	@@:
		359	fld st
		360	fld st
		361	fdivr dword [edi+vec_x]
		362	fstp dword [edi+vec_x]
		363	fdivr dword [edi+vec_y]
		364	fstp dword [edi+vec_y]
		365	fdivr dword [edi+vec_z]
		366	fstp dword [edi+vec_z]
		367	end if
		368	ret
		369	;------------------in: -------------------------
		370	;------------------ esi - pointer to 1st vector
		371	;------------------ edi - pointer to 2nd vector
		372	;------------------out: ------------------------
		373	;------------------ st0 - dot-product
		374	dot_product:
		375	fninit
		376	;if Ext >=SSE3
		377	; movups xmm0,[esi]
		378	; movups xmm1,[edi]
		379	; andps xmm0,[zero_hgst_dd]
		380	; mulps xmm0,xmm1
		381	; haddps xmm0,xmm0
		382	; haddps xmm0,xmm0
		383	; movss [esp-4],xmm0
		384	; fld dword[esp-4]
		385	;else
		386	fld dword [esi+vec_x]
		387	fmul dword [edi+vec_x]
		388	fld dword [esi+vec_y]
		389	fmul dword [edi+vec_y]
		390	fld dword [esi+vec_z]
		391	fmul dword [edi+vec_z]
		392	faddp
		393	faddp
		394	;end if
		395	ret
		396
		397	; DOS version Coded by Mikolaj Felix aka Majuma
		398	; mfelix@polbox.com
		399	; www.majuma.xt.pl
		400	; into FASM translation by Macgub
		401	init_sincos_tab:
		402	.counter equ dword [ebp-4] ; cur angle
		403
		404	push ebp
		405	mov ebp,esp
		406
		407	xor eax,eax
		408	push eax ; init .counter
		409	mov edi,cos_tab
		410	mov esi,sin_tab
		411	mov ecx,256
		412	fninit
		413
		414	fld .counter
		415	@@:
		416	fld st
		417	fsincos
		418	fstp dword [edi]
		419	fstp dword [esi]
		420	; fadd [piD180]
		421	fadd [piD128]
		422	add esi,4
		423	add edi,4
		424	loop @b
		425	ffree st
		426
		427	mov esp,ebp
		428	pop ebp
		429	ret
		430	;------
		431	; esi - offset (pointer) to angles, edi offset to 3x3 matrix
		432	make_rotation_matrix:
		433	.sinx equ dword[ebp-4]
		434	.cosx equ dword[ebp-8]
		435	.siny equ dword[ebp-12]
		436	.cosy equ dword[ebp-16]
		437	.sinz equ dword[ebp-20]
		438	.cosz equ dword[ebp-24]
		439	push ebp
		440	mov ebp,esp
		441	sub esp,24
		442
		443	movzx ebx,word[esi]
		444	shl ebx,2
		445	mov eax,dword[sin_tab+ebx]
		446	mov .sinx,eax
		447	mov edx,dword[cos_tab+ebx]
		448	mov .cosx,edx
		449
		450	movzx ebx,word[esi+2]
		451	shl ebx,2
		452	mov eax,dword[sin_tab+ebx]
		453	mov .siny,eax
		454	mov edx,dword[cos_tab+ebx]
		455	mov .cosy,edx
		456
		457	movzx ebx,word[esi+4]
		458	shl ebx,2
		459	mov eax,dword[sin_tab+ebx]
		460	mov .sinz,eax
		461	mov edx,dword[cos_tab+ebx]
		462	mov .cosz,edx
		463
		464	fninit
		465	fld .cosy
		466	fmul .cosz
		467	fstp dword[edi]
		468
		469	fld .sinx
		470	fmul .siny
		471	fmul .cosz
		472	fld .cosx
		473	fmul .sinz
		474	fchs
		475	faddp
		476	fstp dword[edi+12]
		477
		478	fld .cosx
		479	fmul .siny
		480	fmul .cosz
		481	fld .sinx
		482	fmul .sinz
		483	faddp
		484	fstp dword[edi+24]
		485
		486	fld .cosy
		487	fmul .sinz
		488	fstp dword[edi+4]
		489
		490	fld .sinx
		491	fmul .siny
		492	fmul .sinz
		493	fld .cosx
		494	fmul .cosz
		495	faddp
		496	fstp dword[edi+16]
		497
		498	fld .cosx
		499	fmul .siny
		500	fmul .sinz
		501	fld .sinx
		502	fchs
		503	fmul .cosz
		504	faddp
		505	fstp dword[edi+28]
		506
		507	fld .siny
		508	fchs
		509	fstp dword[edi+8]
		510
		511	fld .cosy
		512	fmul .sinx
		513	fstp dword[edi+20]
		514
		515	fld .cosx
		516	fmul .cosy
		517	fstp dword[edi+32]
		518
		519	mov esp,ebp
		520	pop ebp
		521	ret
		522	;---------------------
		523	; in: esi - ptr to points(normals], each point(normal) coeficient as dword
		524	; edi - ptr to rotated points(normals)
		525	; ebx - ptr to 3x3 (9 dwords, 36 bytes) rotation matrix
		526	; ecx - number of points(normals)
		527	rotary:
		528	if Ext
		529	fninit
		530	.again:
		531
		532	fld dword[esi]
		533	fmul dword[ebx]
		534	fld dword[esi+4]
		535	fmul dword[ebx+12]
		536	faddp
		537	fld dword[esi+8]
		538	fmul dword[ebx+24]
		539	faddp
		540	fstp dword[edi]
		541
		542
		543	fld dword[esi+4]
		544	fmul dword[ebx+16]
		545	fld dword[esi]
		546	fmul dword[ebx+4]
		547	faddp
		548	fld dword[esi+8]
		549	fmul dword[ebx+28]
		550	faddp
		551	fstp dword[edi+4]
		552
		553
		554	fld dword[esi+8]
		555	fmul dword[ebx+32]
		556	fld dword[esi]
		557	fmul dword[ebx+8]
		558	fld dword[esi+4]
		559	fmul dword[ebx+20]
		560	faddp
		561	faddp
		562	fstp dword[edi+8]
		563
		564
		565	add esi,12
		566	add edi,12
		567	loop .again
		568	mov [edi],dword -1
		569	else
		570	; Copyright (C) 1999-2001 Brian Paul
		571	; Copyright (C) Maciej Guba
		572	;---------------------
		573	; in: esi - ptr to points(normals], each point(normal) coeficient as dword
		574	; edi - ptr to rotated points(normals)
		575	; ebx - ptr to 3x3 (9 dwords, 36 bytes) rotation matrix
		576	; ecx - number of points(normals)
		577	;align 32
		578	movups xmm4,[ebx]
		579	; lddqu xmm4,[ebx] ; I tried sse3 :D
		580	movups xmm5,[ebx+12]
		581	movups xmm6,[ebx+24]
		582	;align 32
		583	.again:
		584	movss xmm0,dword[esi]
		585	shufps xmm0,xmm0,0
		586	mulps xmm0,xmm4
		587
		588	movss xmm1,dword[esi+4]
		589	shufps xmm1,xmm1,0
		590	mulps xmm1,xmm5
		591
		592	movss xmm2,dword[esi+8]
		593	shufps xmm2,xmm2,0
		594	mulps xmm2,xmm6
		595
		596	addps xmm0,xmm1
		597	addps xmm0,xmm2
		598
		599	movups [edi],xmm0
		600
		601	add esi,12
		602	add edi,12
9740	macgub	603	; dec ecx
		604	; jne .again
		605	loop .again
8719	leency	606	mov [edi],dword -1
		607	end if
		608	ret
		609	;----------------------------------------------
		610	; esi - pointer to 3x3 matrix
		611	add_scale_to_matrix:
9512	IgorA	612	if Ext>SSE
		613	movss xmm0,[rsscale]
		614	shufps xmm0,xmm0,0
		615	movups xmm1,[esi]
		616	movups xmm2,[esi+16]
		617	movss xmm3,[esi+32]
		618	mulps xmm1,xmm0
		619	mulps xmm2,xmm0
		620	mulss xmm3,xmm0
		621	movups [esi],xmm1
		622	movups [esi+16],xmm2
		623	movss [esi+32],xmm3
		624	else
8719	leency	625	fninit
		626	fld [rsscale]
		627	fld dword[esi] ;-----
		628	fmul st,st1
		629	fstp dword[esi]
		630	fld dword[esi+12] ; x scale
		631	fmul st,st1
		632	fstp dword[esi+12]
		633	fld dword[esi+24]
		634	fmul st,st1
		635	fstp dword[esi+24] ;------
		636
		637	fld dword[esi+4] ;-----
		638	fmul st,st1
		639	fstp dword[esi+4]
		640	fld dword[esi+16] ; y scale
		641	fmul st,st1
		642	fstp dword[esi+16]
		643	fld dword[esi+28]
		644	fmul st,st1
		645	fstp dword[esi+28] ;------
		646
		647
		648	fld dword[esi+8] ;-----
		649	fmul st,st1
		650	fstp dword[esi+8]
		651	fld dword[esi+20] ; z scale
		652	fmul st,st1
		653	fstp dword[esi+20]
		654	fld dword[esi+32]
		655	fmulp st1,st
		656	fstp dword[esi+32] ;------
9512	IgorA	657	end if
8719	leency	658	ret
		659
		660	;in esi - offset to 3d points (point as 3 dwords float)
		661	; edi - offset to 2d points ( as 3 words integer)
		662	; ecx - number of points
		663	translate_points: ; just convert into integer; z coord still needed
9512	IgorA	664	if Ext < SSE
8719	leency	665	fninit
9512	IgorA	666	else
		667	; movaps xmm1,[vect_x]
		668	end if
		669
8719	leency	670	.again:
9740	macgub	671	if 0
8719	leency	672	fld dword[esi+8]
		673	; fmul [rsscale]
		674	fist word[edi+4]
		675
		676	fisub [zobs]
		677	fchs
		678
		679	fld dword[esi]
		680	; fmul [rsscale]
		681	fisub [xobs]
		682	fimul [zobs]
		683	fdiv st0,st1
		684
		685	fiadd [xobs]
		686	fiadd [vect_x]
		687	fistp word[edi]
		688
		689	fld dword[esi+4]
		690	; fmul [rsscale]
		691	fisub [yobs]
		692	fimul [zobs]
		693	fdivrp ; st0,st1
		694
		695	fiadd [yobs]
		696	fiadd [vect_y]
		697	fistp word[edi+2]
		698	end if
9740	macgub	699	if Ext>=SSE2
9512	IgorA	700	movups xmm0,[esi]
		701	cvtps2dq xmm0,xmm0
		702	packssdw xmm0,xmm0
		703	paddw xmm0,[vect_x]
		704	movd [edi],xmm0
		705	; psrldq xmm0,4
		706	; movd eax,xmm0
		707	pextrw eax,xmm0,6
		708	mov [edi+4],ax
		709	else
		710
8719	leency	711	; cvtps2dq xmm0,xmm0
		712	; packsdw xmm0,xmm0
		713	; movq [edi]
		714	fld dword[esi]
9237	leency	715	fiadd word[vect_x]
8719	leency	716	fistp word[edi]
		717	fld dword[esi+4]
		718	fiadd [vect_y]
		719	fistp word[edi+2]
		720	fld dword[esi+8]
		721	fistp word[edi+4]
9512	IgorA	722	end if
		723
8719	leency	724	add esi,12
		725	add edi,6
9512	IgorA	726	loop .again
8719	leency	727
		728	ret

Subversion Repositories Kolibri OS

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