ShuriZma
/
pcsx2
mirror of https://github.com/PCSX2/pcsx2.git
1
0
Fork 0
Code Issues Packages Projects Releases Wiki Activity

a few small optimizations and changes for the VU recs. 

git-svn-id: http://pcsx2-playground.googlecode.com/svn/trunk@15 a6443dda-0b58-4228-96e9-037be469359c
This commit is contained in:
cottonvibes 2008-08-13 07:48:58 +00:00 committed by Gregory Hainaut
parent 7bf97e1af1
commit c667ffe343
1 changed files with 18 additions and 22 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

40
pcsx2/x86/iVUmicro.c
View File

@ -1263,7 +1263,11 @@ void vuFloat(int regd, int XYZW) {
// Clamps infinities to max/min non-infinity number (uses a temp reg)
void vuFloat2(int regd, int regTemp, int XYZW) {
if( CHECK_OVERFLOW ) {
if (XYZW != 0xf) { // here we use a temp reg because not all xyzw are being modified
if (XYZW == 8) {
SSE_MINSS_M32_to_XMM(regd, (uptr)g_maxvals);
SSE_MAXSS_M32_to_XMM(regd, (uptr)g_minvals);
}
else if (XYZW != 0xf) { // here we use a temp reg because not all xyzw are being modified
SSE_MOVAPS_XMM_to_XMM(regTemp, regd);
SSE_MINPS_M128_to_XMM(regTemp, (uptr)g_maxvals);
SSE_MAXPS_M128_to_XMM(regTemp, (uptr)g_minvals);
@ -5309,7 +5313,7 @@ void vuSqSumXYZ(int regd, int regs, int regtemp)
else
{
SSE_MOVAPS_XMM_to_XMM(regtemp, regs);
SSE_MULPS_XMM_to_XMM(regtemp, regtemp);
SSE_MULPS_XMM_to_XMM(regtemp, regtemp); // xyzw ^ 2
if( cpucaps.hasStreamingSIMD3Extensions ) {
SSE3_HADDPS_XMM_to_XMM(regd, regtemp);
@ -5318,10 +5322,10 @@ void vuSqSumXYZ(int regd, int regs, int regtemp)
}
else {
SSE_MOVSS_XMM_to_XMM(regd, regtemp);
SSE_SHUFPS_XMM_to_XMM(regtemp, regtemp, 0xE1);
SSE_ADDSS_XMM_to_XMM(regd, regtemp);
SSE_SHUFPS_XMM_to_XMM(regtemp, regtemp, 0xD2);
SSE_ADDSS_XMM_to_XMM(regd, regtemp);
SSE_SHUFPS_XMM_to_XMM(regtemp, regtemp, 0xE1); // wzyx -> wzxy
SSE_ADDSS_XMM_to_XMM(regd, regtemp); // x ^ 2 + y ^ 2
SSE_SHUFPS_XMM_to_XMM(regtemp, regtemp, 0xD2); // wzxy -> wxyz
SSE_ADDSS_XMM_to_XMM(regd, regtemp); // x ^ 2 + y ^ 2 + z ^ 2
//SSE_SHUFPS_XMM_to_XMM(regtemp, regtemp, 0xC6);
}
}
@ -5356,9 +5360,7 @@ void recVUMI_ESADD( VURegs *VU, int info)
}
}
CheckForOverflowSS_(EEREC_TEMP, EEREC_D);
//SSE_MINSS_M32_to_XMM(EEREC_TEMP, (uptr)&g_maxvals[0]);
//SSE_MAXSS_M32_to_XMM(EEREC_TEMP, (uptr)&g_minvals[0]);
vuFloat2(EEREC_TEMP, EEREC_D, 0x8);
SSE_MOVSS_XMM_to_M32(VU_VI_ADDR(REG_P, 0), EEREC_TEMP);
}
@ -5395,9 +5397,7 @@ void recVUMI_ERSADD( VURegs *VU, int info )
// don't use RCPSS (very bad precision)
SSE_DIVSS_XMM_to_XMM(EEREC_TEMP, EEREC_D);
CheckForOverflowSS_(EEREC_TEMP, EEREC_D);
//SSE_MINSS_M32_to_XMM(EEREC_TEMP, (uptr)&g_maxvals[0]);
//SSE_MAXSS_M32_to_XMM(EEREC_TEMP, (uptr)&g_minvals[0]);
vuFloat2(EEREC_TEMP, EEREC_D, 0x8);
SSE_MOVSS_XMM_to_M32(VU_VI_ADDR(REG_P, 0), EEREC_TEMP);
}
@ -5414,17 +5414,13 @@ void recVUMI_ELENG( VURegs *VU, int info )
void recVUMI_ERLENG( VURegs *VU, int info )
{
assert( VU == &VU1 );
vuSqSumXYZ(EEREC_D, EEREC_S, EEREC_TEMP); //Dont want to use EEREC_D incase it overwrites something
vuSqSumXYZ(EEREC_D, EEREC_S, EEREC_TEMP);
//SysPrintf("ERLENG\n");
SSE_SQRTSS_XMM_to_XMM(EEREC_TEMP, EEREC_D);
SSE_MOVSS_XMM_to_M32(VU_VI_ADDR(REG_P, 0), EEREC_TEMP);
SSE_MOVSS_M32_to_XMM(EEREC_TEMP, (uptr)&VU->VF[0].UL[3]);
SSE_DIVSS_M32_to_XMM(EEREC_TEMP, VU_VI_ADDR(REG_P, 0));
//SSE_RSQRTSS_XMM_to_XMM(EEREC_TEMP, EEREC_D);
//CheckForOverflowSS_(EEREC_TEMP, EEREC_D);
//SSE_MINSS_M32_to_XMM(EEREC_TEMP, (uptr)&g_maxvals[0]);
//SSE_MAXSS_M32_to_XMM(EEREC_TEMP, (uptr)&g_minvals[0]);
SSE_SQRTSS_XMM_to_XMM(EEREC_TEMP, EEREC_D); // sqrt(x^2 + y^2 + z^2)
SSE_MOVSS_XMM_to_XMM(EEREC_D, EEREC_TEMP); // d <- sqrt(x^2 + y^2 + z^2)
SSE_MOVSS_M32_to_XMM(EEREC_TEMP, (uptr)&VU->VF[0].UL[3]); // temp <- 1
SSE_DIVSS_XMM_to_XMM(EEREC_TEMP, EEREC_D); // temp = 1 / sqrt(x^2 + y^2 + z^2)
vuFloat2(EEREC_TEMP, EEREC_D, 0x8);
SSE_MOVSS_XMM_to_M32(VU_VI_ADDR(REG_P, 0), EEREC_TEMP);
}