project64/Source/Glide64/3dmathSIMD.asm.cpp

493 lines
14 KiB
C++

;/*
;* Glide64 - Glide video plugin for Nintendo 64 emulators.
;*
;* This program is free software; you can redistribute it and/or modify
;* it under the terms of the GNU General Public License as published by
;* the Free Software Foundation; either version 2 of the License, or
;* any later version.
;*
;* This program is distributed in the hope that it will be useful,
;* but WITHOUT ANY WARRANTY; without even the implied warranty of
;* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
;* GNU General Public License for more details.
;*
;* You should have received a copy of the GNU General Public License
;* along with this program; if not, write to the Free Software
;* Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA
;
;****************************************************************
;
; Glide64 - Glide Plugin for Nintendo 64 emulators
; Project started on December 29th, 2001
;
; Authors:
; Dave2001, original author, founded the project in 2001, left it in 2002
; Gugaman, joined the project in 2002, left it in 2002
; Sergey 'Gonetz' Lipski, joined the project in 2002, main author since fall of 2002
; Hiroshi 'KoolSmoky' Morii, joined the project in 2007
;
;****************************************************************
;
; To modify Glide64:
; * Write your name and (optional)email, commented by your work, so I know who did it, and so that you can find which parts you modified when it comes time to send it to me.
; * Do NOT send me the whole project or file that you modified. Take out your modified code sections, and tell me where to put them. If people sent the whole thing, I would have many different versions, but no idea how to combine them all.
;
;****************************************************************
%include "inc/c32.mac"
segment .text
*/
extern "C" void __declspec(naked) DetectSIMD(int func, int * iedx, int * iecx)
{
_asm {
push ebp
mov ebp,esp
mov eax,[func]
cpuid
mov eax,[iedx]
mov [eax],edx
mov eax,[iecx]
mov [eax],ecx
leave
ret
}
}
/****************************************************************
;
; ******** SSE ********
;
;****************************************************************/
extern "C" void __declspec(naked) MulMatricesSSE(float m1[4][4],float m2[4][4],float r[4][4])
{
__asm
{
push ebp
mov ebp,esp
mov eax,[r]
mov ecx,[m1]
mov edx,[m2]
movaps xmm0,[edx]
movaps xmm1,[edx+16]
movaps xmm2,[edx+32]
movaps xmm3,[edx+48]
; r[0][0],r[0][1],r[0][2],r[0][3]
movaps xmm4,[ecx]
movaps xmm5,xmm4
movaps xmm6,xmm4
movaps xmm7,xmm4
shufps xmm4,xmm4,00000000b
shufps xmm5,xmm5,01010101b
shufps xmm6,xmm6,10101010b
shufps xmm7,xmm7,11111111b
mulps xmm4,xmm0
mulps xmm5,xmm1
mulps xmm6,xmm2
mulps xmm7,xmm3
addps xmm4,xmm5
addps xmm4,xmm6
addps xmm4,xmm7
movaps [eax],xmm4
; r[1][0],r[1][1],r[1][2],r[1][3]
movaps xmm4,[ecx+16]
movaps xmm5,xmm4
movaps xmm6,xmm4
movaps xmm7,xmm4
shufps xmm4,xmm4,00000000b
shufps xmm5,xmm5,01010101b
shufps xmm6,xmm6,10101010b
shufps xmm7,xmm7,11111111b
mulps xmm4,xmm0
mulps xmm5,xmm1
mulps xmm6,xmm2
mulps xmm7,xmm3
addps xmm4,xmm5
addps xmm4,xmm6
addps xmm4,xmm7
movaps [eax+16],xmm4
; r[2][0],r[2][1],r[2][2],r[2][3]
movaps xmm4,[ecx+32]
movaps xmm5,xmm4
movaps xmm6,xmm4
movaps xmm7,xmm4
shufps xmm4,xmm4,00000000b
shufps xmm5,xmm5,01010101b
shufps xmm6,xmm6,10101010b
shufps xmm7,xmm7,11111111b
mulps xmm4,xmm0
mulps xmm5,xmm1
mulps xmm6,xmm2
mulps xmm7,xmm3
addps xmm4,xmm5
addps xmm4,xmm6
addps xmm4,xmm7
movaps [eax+32],xmm4
; r[3][0],r[3][1],r[3][2],r[3][3]
movaps xmm4,[ecx+48]
movaps xmm5,xmm4
movaps xmm6,xmm4
movaps xmm7,xmm4
shufps xmm4,xmm4,00000000b
shufps xmm5,xmm5,01010101b
shufps xmm6,xmm6,10101010b
shufps xmm7,xmm7,11111111b
mulps xmm4,xmm0
mulps xmm5,xmm1
mulps xmm6,xmm2
mulps xmm7,xmm3
addps xmm4,xmm5
addps xmm4,xmm6
addps xmm4,xmm7
movaps [eax+48],xmm4
leave
ret
}
}
extern "C" void __declspec(naked) NormalizeVectorSSE (float *v)
{
_asm
{
push ebp
mov ebp,esp
mov edx, [v]
movaps xmm0, [edx] ; x y z 0
movaps xmm2, xmm0 ; x y z 0
mulps xmm0, xmm0 ; x*x y*y z*z 0
movaps xmm1, xmm0 ; x*x y*y z*z 0
shufps xmm0, xmm1, 0x4e ; z*z 0 x*x y*y
addps xmm0, xmm1 ; x*x+z*z y*y z*z+x*x y*y
movaps xmm1, xmm0 ; x*x+z*z y*y z*z+x*x y*y
shufps xmm1, xmm1, 0x11 ; y*y z*z+x*x y*y z*z+x*x
addps xmm0, xmm1 ; x*x+z*z+y*y
rsqrtps xmm0, xmm0 ; 1.0/sqrt(x*x+z*z+y*y)
mulps xmm2, xmm0 ; x/sqrt(x*x+z*z+y*y) y/sqrt(x*x+z*z+y*y) z/sqrt(x*x+z*z+y*y) 0
movaps [edx], xmm2
leave
ret
}
}
/*****************************************************************
;
; ******** SSE3 ********
;
;****************************************************************/
float __declspec(naked) DotProductSSE3(register float *v1, register float *v2)
{
_asm {
push ebp
mov ebp,esp
mov eax,[v1]
mov edx,[v2]
movaps xmm0, [eax]
mulps xmm0, [edx]
haddps xmm0, xmm0
haddps xmm0, xmm0
; movss eax, xmm0
leave
ret
}
}
/****************************************************************
;
; ******** 3DNOW ********
;
;****************************************************************/
extern "C" void __declspec(naked) TransformVector3DNOW(float *src, float *dst, float mat[4][4])
{
_asm {
push ebp
mov ebp,esp
femms
mov ecx,[src]
mov eax,[dst]
mov edx,[mat]
movq mm0,[ecx] ; src[1] src[0]
movd mm2,[ecx+8] ; 0 src[2]
movq mm1,mm0 ; src[1] src[0]
punpckldq mm0,mm0 ; src[0] src[0]
punpckhdq mm1,mm1 ; src[1] src[1]
punpckldq mm2,mm2 ; src[2] src[2]
movq mm3,mm0 ; src[0] src[0]
movq mm4,mm1 ; src[1] src[1]
movq mm5,mm2 ; src[2] src[2]
pfmul mm0,[edx] ; src[0]*mat[0][1] src[0]*mat[0][0]
pfmul mm3,[edx+8] ; 0 src[0]*mat[0][2]
pfmul mm1,[edx+16] ; src[1]*mat[1][1] src[1]*mat[1][0]
pfmul mm4,[edx+24] ; 0 src[1]*mat[1][2]
pfmul mm2,[edx+32] ; src[2]*mat[2][1] src[2]*mat[2][0]
pfmul mm5,[edx+40] ; 0 src[2]*mat[2][2]
pfadd mm0,mm1 ; src[0]*mat[0][1]+src[1]*mat[1][1] src[0]*mat[0][0]+src[1]*mat[1][0]
pfadd mm3,mm4 ; 0 src[0]*mat[0][2]+src[1]*mat[1][2]
pfadd mm0,mm2 ; src[0]*mat[0][1]+src[1]*mat[1][1]+src[2]*mat[2][1] src[0]*mat[0][0]+src[1]*mat[1][0]+src[2]*mat[2][0]
pfadd mm3,mm5 ; 0 src[0]*mat[0][2]+src[1]*mat[1][2]+src[2]*mat[2][2]
movq [eax],mm0 ; mat[0][1]*src[0]+mat[1][1]*src[1]+mat[2][1]*src[2] mat[0][0]*src[0]+mat[1][0]*src[1]+mat[2][0]*src[2]
movd [eax+8],mm3 ; mat[0][2]*src[0]+mat[1][2]*src[1]+mat[2][2]*src[2]
femms
leave
ret
}
}
extern "C" void __declspec(naked) InverseTransformVector3DNOW(float *src, float *dst, float mat[4][4])
{
_asm {
push ebp
mov ebp,esp
femms
mov ecx,[src]
mov eax,[dst]
mov edx,[mat]
movq mm0,[ecx] ; src[1] src[0]
movd mm4,[ecx+8] ; 0 src[2]
movq mm1,mm0 ; src[1] src[0]
pfmul mm0,[edx] ; src[1]*mat[0][1] src[0]*mat[0][0]
movq mm5,mm4 ; 0 src[2]
pfmul mm4,[edx+8] ; 0 src[2]*mat[0][2]
movq mm2,mm1 ; src[1] src[0]
pfmul mm1,[edx+16] ; src[1]*mat[1][1] src[0]*mat[1][0]
movq mm6,mm5 ; 0 src[2]
pfmul mm5,[edx+24] ; 0 src[2]*mat[1][2]
movq mm3,mm2 ; src[1] src[0]
pfmul mm2,[edx+32] ; src[1]*mat[2][1] src[0]*mat[2][0]
movq mm7,mm6 ; 0 src[2]
pfmul mm6,[edx+40] ; 0 src[2]*mat[2][2]
pfacc mm0,mm4 ; src[2]*mat[0][2] src[1]*mat[0][1]+src[0]*mat[0][0]
pfacc mm1,mm5 ; src[2]*mat[1][2] src[1]*mat[1][1]+src[0]*mat[1][0]
pfacc mm2,mm6 ; src[2]*mat[2][2] src[1]*mat[2][1]+src[0]*mat[2][0]
pfacc mm0,mm1 ; src[2]*mat[1][2]+src[1]*mat[1][1]+src[0]*mat[1][0] src[2]*mat[0][2]+src[1]*mat[0][1]+src[0]*mat[0][0]
pfacc mm2,mm3 ; 0 src[2]*mat[2][2]+src[1]*mat[2][1]+src[0]*mat[2][0]
movq [eax],mm0 ; mat[1][0]*src[0]+mat[1][1]*src[1]+mat[1][2]*src[2] mat[0][0]*src[0]+mat[0][1]*src[1]+mat[0][2]*src[2]
movd [eax+8],mm2 ; mat[2][0]*src[0]+mat[2][1]*src[1]+mat[2][2]*src[2]
femms
leave
ret
}
}
extern "C" void __declspec(naked) MulMatrices3DNOW(float m1[4][4],float m2[4][4],float r[4][4])
{
_asm {
push ebp
mov ebp,esp
femms
mov ecx,[m1]
mov eax,[r]
mov edx,[m2]
movq mm0,[ecx]
movq mm1,[ecx+8]
movq mm4,[edx]
punpckhdq mm2,mm0
movq mm5,[edx+16]
punpckhdq mm3,mm1
movq mm6,[edx+32]
punpckldq mm0,mm0
punpckldq mm1,mm1
pfmul mm4,mm0
punpckhdq mm2,mm2
pfmul mm0,[edx+8]
movq mm7,[edx+48]
pfmul mm5,mm2
punpckhdq mm3,mm3
pfmul mm2,[edx+24]
pfmul mm6,mm1
pfadd mm5,mm4
pfmul mm1,[edx+40]
pfadd mm2,mm0
pfmul mm7,mm3
pfadd mm6,mm5
pfmul mm3,[edx+56]
pfadd mm2,mm1
pfadd mm7,mm6
movq mm0,[ecx+16]
pfadd mm3,mm2
movq mm1,[ecx+24]
movq [eax],mm7
movq mm4,[edx]
movq [eax+8],mm3
punpckhdq mm2,mm0
movq mm5,[edx+16]
punpckhdq mm3,mm1
movq mm6,[edx+32]
punpckldq mm0,mm0
punpckldq mm1,mm1
pfmul mm4,mm0
punpckhdq mm2,mm2
pfmul mm0,[edx+8]
movq mm7,[edx+48]
pfmul mm5,mm2
punpckhdq mm3,mm3
pfmul mm2,[edx+24]
pfmul mm6,mm1
pfadd mm5,mm4
pfmul mm1,[edx+40]
pfadd mm2,mm0
pfmul mm7,mm3
pfadd mm6,mm5
pfmul mm3,[edx+56]
pfadd mm2,mm1
pfadd mm7,mm6
movq mm0,[ecx+32]
pfadd mm3,mm2
movq mm1,[ecx+40]
movq [eax+16],mm7
movq mm4,[edx]
movq [eax+24],mm3
punpckhdq mm2,mm0
movq mm5,[edx+16]
punpckhdq mm3,mm1
movq mm6,[edx+32]
punpckldq mm0,mm0
punpckldq mm1,mm1
pfmul mm4,mm0
punpckhdq mm2,mm2
pfmul mm0,[edx+8]
movq mm7,[edx+48]
pfmul mm5,mm2
punpckhdq mm3,mm3
pfmul mm2,[edx+24]
pfmul mm6,mm1
pfadd mm5,mm4
pfmul mm1,[edx+40]
pfadd mm2,mm0
pfmul mm7,mm3
pfadd mm6,mm5
pfmul mm3,[edx+56]
pfadd mm2,mm1
pfadd mm7,mm6
movq mm0,[ecx+48]
pfadd mm3,mm2
movq mm1,[ecx+56]
movq [eax+32],mm7
movq mm4,[edx]
movq [eax+40],mm3
punpckhdq mm2,mm0
movq mm5,[edx+16]
punpckhdq mm3,mm1
movq mm6,[edx+32]
punpckldq mm0,mm0
punpckldq mm1,mm1
pfmul mm4,mm0
punpckhdq mm2,mm2
pfmul mm0,[edx+8]
movq mm7,[edx+48]
pfmul mm5,mm2
punpckhdq mm3,mm3
pfmul mm2,[edx+24]
pfmul mm6,mm1
pfadd mm5,mm4
pfmul mm1,[edx+40]
pfadd mm2,mm0
pfmul mm7,mm3
pfadd mm6,mm5
pfmul mm3,[edx+56]
pfadd mm2,mm1
pfadd mm7,mm6
pfadd mm3,mm2
movq [eax+48],mm7
movq [eax+56],mm3
femms
leave
ret
}
}
extern "C" float __declspec(naked) DotProduct3DNOW(register float *v1, register float *v2)
{
_asm {
push ebp
mov ebp,esp
femms
mov edx,[v1]
mov eax,[v2]
movq mm0,[edx]
movq mm3,[eax]
pfmul mm0,mm3
movq mm2,[edx+8]
movq mm1,[eax+8]
pfacc mm0,mm0
pfmul mm1,mm2
pfadd mm0,mm1
movd eax,mm0
femms
leave
ret
}
}
extern "C" void __declspec(naked) NormalizeVector3DNOW(float *v)
{
_asm {
push ebp
mov ebp,esp
femms
mov edx,[v]
movq mm0,[edx]
movq mm3,[edx+8]
movq mm1,mm0
movq mm2,mm3
pfmul mm0,mm0
pfmul mm3,mm3
pfacc mm0,mm0
pfadd mm0,mm3
;movq mm4,mm0 ; prepare for 24bit precision
;punpckldq mm4,mm4 ; prepare for 24bit precision
pfrsqrt mm0,mm0 ; 15bit precision 1/sqrtf(v)
;movq mm3,mm0
;pfmul mm0,mm0
;pfrsqit1 mm0,mm4
;pfrcpit2 mm0,mm3 ; 24bit precision 1/sqrtf(v)
pfmul mm1,mm0
pfmul mm2,mm0
movq [edx],mm1
movq [edx+8],mm2
femms
leave
ret
}
}