piwalker/dolphin
1
0
Fork 0
mirror of https://github.com/dolphin-emu/dolphin.git synced 2025-07-21 05:09:34 -06:00
Code Issues Packages Projects Releases Wiki Activity

TextureDecoder.cpp: new SSE2 optimized GX_TF_I8 decoder. Probably not ultimately optimal SSE2 code, but provably better (on my machine) than the memset version. Tested with __rdtsc counts in an independent project. I get about 6-7 FPS more on average during the intro movie playback in Mario Kart Wii. Hope this compiles for GCC okay.

Browse Source 
TextureDecoder.cpp: merged two functionally identical decode5A3RGBA and decode5A3rgba methods.
OpcodeDecoding.cpp and DLCache.cpp: optimization for GX_LOAD_XF_REG. The PSUHFB solution sounds better for SSSE3, but this is a small win for the default case.

git-svn-id: https://dolphin-emu.googlecode.com/svn/trunk@6692 8ced0084-cf51-0410-be5f-012b33b47a6e
This commit is contained in:
james.jdunne
2010-12-30 19:17:08 +00:00
parent 6cf9b3688d
commit b038df64bf
5 changed files with 164 additions and 66 deletions
Download Patch File Download Diff File Expand all files Collapse all files

26
Source/Core/VideoCommon/Src/DataReader.h
View File

@ -64,6 +64,32 @@ __forceinline u32 DataReadU32()
return tmp;
}
template<unsigned int N>
void DataReadU32xN(u32 *bufx16)
{
memcpy(bufx16, g_pVideoData, sizeof(u32) * N);
if (N >= 1) bufx16[0] = Common::swap32(bufx16[0]);
if (N >= 2) bufx16[1] = Common::swap32(bufx16[1]);
if (N >= 3) bufx16[2] = Common::swap32(bufx16[2]);
if (N >= 4) bufx16[3] = Common::swap32(bufx16[3]);
if (N >= 5) bufx16[4] = Common::swap32(bufx16[4]);
if (N >= 6) bufx16[5] = Common::swap32(bufx16[5]);
if (N >= 7) bufx16[6] = Common::swap32(bufx16[6]);
if (N >= 8) bufx16[7] = Common::swap32(bufx16[7]);
if (N >= 9) bufx16[8] = Common::swap32(bufx16[8]);
if (N >= 10) bufx16[9] = Common::swap32(bufx16[9]);
if (N >= 11) bufx16[10] = Common::swap32(bufx16[10]);
if (N >= 12) bufx16[11] = Common::swap32(bufx16[11]);
if (N >= 13) bufx16[12] = Common::swap32(bufx16[12]);
if (N >= 14) bufx16[13] = Common::swap32(bufx16[13]);
if (N >= 15) bufx16[14] = Common::swap32(bufx16[14]);
if (N >= 16) bufx16[15] = Common::swap32(bufx16[15]);
g_pVideoData += (sizeof(u32) * N);
}
typedef void (*DataReadU32xNfunc)(u32 *buf);
extern DataReadU32xNfunc DataReadU32xFuncs[16];
__forceinline u32 DataReadU32Unswapped()
{
u32 tmp = *(u32*)g_pVideoData;