piwalker/dolphin
1
0
Fork 0
mirror of https://github.com/dolphin-emu/dolphin.git synced 2025-07-23 06:09:50 -06:00
Code Issues Packages Projects Releases Wiki Activity

sometimes to advance you have to make a step back.

Browse Source 
use plain vertex arrays instead of VBOs to render in Opengl plugin as the nature of the data make VBOs slower. This must bring, depending on the implementation, a good speedup in opengl.
in my system now opengl and d3d9 have a difference of 1 to 5 fps depending of the game.
some cleanup and a little work pointing to future improvements in the way of rendering.
please test and check for any errors.

git-svn-id: https://dolphin-emu.googlecode.com/svn/trunk@6139 8ced0084-cf51-0410-be5f-012b33b47a6e
This commit is contained in:
Rodolfo Osvaldo Bogado
2010-08-28 15:09:42 +00:00
parent eb40f765c2
commit 9b0357b5e2
18 changed files with 414 additions and 424 deletions
Download Patch File Download Diff File Expand all files Collapse all files

116
Source/Core/Common/Src/Hash.cpp
View File

@ -119,3 +119,119 @@ u32 HashEctor(const u8* ptr, int length)
return(crc);
}
#ifdef _M_X64
u64 GetHash64(const u8 *src, int len, u32 samples)
{
const u64 m = 0xc6a4a7935bd1e995;
const int r = 47;
u64 h = len * m;
u32 Step = (len/8);
const u64 * data = (const u64 *)src;
const u64 * end = data + Step;
if(samples == 0) samples = Step;
Step = Step / samples;
if(Step < 1) Step = 1;
while(data < end)
{
u64 k = data[0];
data+=Step;
k *= m;
k ^= k >> r;
k *= m;
h ^= k;
h *= m;
}
const u8 * data2 = (const u8*)end;
switch(len & 7)
{
case 7: h ^= u64(data2[6]) << 48;
case 6: h ^= u64(data2[5]) << 40;
case 5: h ^= u64(data2[4]) << 32;
case 4: h ^= u64(data2[3]) << 24;
case 3: h ^= u64(data2[2]) << 16;
case 2: h ^= u64(data2[1]) << 8;
case 1: h ^= u64(data2[0]);
h *= m;
};
h ^= h >> r;
h *= m;
h ^= h >> r;
return h;
}
#else
u64 GetHash64(const u8 *src, int len, u32 samples)
{
const u32 m = 0x5bd1e995;
const int r = 24;
u32 h1 = len;
u32 h2 = 0;
u32 Step = (len / 4);
const u32 * data = (const u32 *)src;
const u32 * end = data + Step;
const u8 * uEnd = (const u8 *)end;
if(samples == 0) samples = Step;
Step = Step / samples;
if(Step < 2) Step = 2;
while(data < end)
{
u32 k1 = data[0];
k1 *= m;
k1 ^= k1 >> r;
k1 *= m;
h1 *= m;
h1 ^= k1;
u32 k2 = data[1];
k2 *= m;
k2 ^= k2 >> r;
k2 *= m;
h2 *= m;
h2 ^= k2;
data+=Step;
}
if((len & 7) > 3)
{
u32 k1 = *(end - 1);
k1 *= m;
k1 ^= k1 >> r;
k1 *= m;
h1 *= m;
h1 ^= k1;
len -= 4;
}
switch(len & 3)
{
case 3: h2 ^= uEnd[2] << 16;
case 2: h2 ^= uEnd[1] << 8;
case 1: h2 ^= uEnd[0];
h2 *= m;
};
h1 ^= h2 >> 18; h1 *= m;
h2 ^= h1 >> 22; h2 *= m;
h1 ^= h2 >> 17; h1 *= m;
h2 ^= h1 >> 19; h2 *= m;
u64 h = h1;
h = (h << 32) | h2;
return h;
}
#endif

2
Source/Core/Common/Src/Hash.h
View File

@ -24,5 +24,5 @@ u32 HashFletcher(const u8* data_u8, size_t length); // FAST. Length & 1 == 0.
u32 HashAdler32(const u8* data, size_t len); // Fairly accurate, slightly slower
u32 HashFNV(const u8* ptr, int length); // Another fast and decent hash
u32 HashEctor(const u8* ptr, int length); // JUNK. DO NOT USE FOR NEW THINGS
u64 GetHash64(const u8 *src, int len, u32 samples);
#endif // _HASH_H_

2
Source/Core/Common/Src/LinearDiskCache.cpp
View File

@ -22,7 +22,7 @@ static const char ID[4] = {'D', 'C', 'A', 'C'};
// Update this to the current SVN revision every time you change shader generation code.
// We don't automatically get this from SVN_REV because that would mean regenerating the
// shader cache for every revision, graphics-related or not, which is simply annoying.
const int version = 6124;
const int version = 6139;
LinearDiskCache::LinearDiskCache()
: file_(NULL), num_entries_(0) {

1
Source/Core/VideoCommon/Src/NativeVertexWriter.cpp
View File

@ -22,6 +22,7 @@ namespace VertexManager
{
u8* s_pCurBufferPointer = NULL;
u8* s_pBaseBufferPointer = NULL;
}

1
Source/Core/VideoCommon/Src/NativeVertexWriter.h
View File

@ -29,6 +29,7 @@ int GetRemainingVertices(int primitive); // remaining number of vertices that ca
// TODO: move, rename.
extern u8* s_pCurBufferPointer;
extern u8* s_pBaseBufferPointer;
}

4
Source/Core/VideoCommon/Src/OpcodeDecoding.cpp
View File

@ -233,12 +233,12 @@ static void Decode()
{
u32 Cmd2 = DataReadU32();
int transfer_size = ((Cmd2 >> 16) & 15) + 1;
u32 address = Cmd2 & 0xFFFF;
u32 xf_address = Cmd2 & 0xFFFF;
// TODO - speed this up. pshufb?
u32 data_buffer[16];
for (int i = 0; i < transfer_size; i++)
data_buffer[i] = DataReadU32();
LoadXFReg(transfer_size, address, data_buffer);
LoadXFReg(transfer_size, xf_address, data_buffer);
INCSTAT(stats.thisFrame.numXFLoads);
}
break;

2
Source/Core/VideoCommon/Src/OpcodeDecoding.h
View File

@ -48,5 +48,5 @@
void OpcodeDecoder_Init();
void OpcodeDecoder_Shutdown();
void OpcodeDecoder_Run(bool skipped_frame);
void ExecuteDisplayList(u32 address, u32 size);
#endif // _OPCODE_DECODING_H

117
Source/Core/VideoCommon/Src/TextureDecoder.cpp
View File

@ -20,7 +20,6 @@
#include "CPUDetect.h"
#include "TextureDecoder.h"
#include "OpenCL.h"
#if defined(HAVE_OPENCL) && HAVE_OPENCL
#include "OpenCL/OCLTextureDecoder.h"
@ -96,122 +95,6 @@ int TexDecoder_GetTextureSizeInBytes(int width, int height, int format)
return (width * height * TexDecoder_GetTexelSizeInNibbles(format)) / 2;
}
#ifdef _M_X64
u64 TexDecoder_GetHash64(const u8 *src, int len, u32 samples)
{
const u64 m = 0xc6a4a7935bd1e995;
const int r = 47;
u64 h = len * m;
u32 Step = (len/8);
const u64 * data = (const u64 *)src;
const u64 * end = data + Step;
if(samples == 0) samples = Step;
Step = Step / samples;
if(Step < 1) Step = 1;
while(data < end)
{
u64 k = data[0];
data+=Step;
k *= m;
k ^= k >> r;
k *= m;
h ^= k;
h *= m;
}
const u8 * data2 = (const u8*)end;
switch(len & 7)
{
case 7: h ^= u64(data2[6]) << 48;
case 6: h ^= u64(data2[5]) << 40;
case 5: h ^= u64(data2[4]) << 32;
case 4: h ^= u64(data2[3]) << 24;
case 3: h ^= u64(data2[2]) << 16;
case 2: h ^= u64(data2[1]) << 8;
case 1: h ^= u64(data2[0]);
h *= m;
};
h ^= h >> r;
h *= m;
h ^= h >> r;
return h;
}
#else
u64 TexDecoder_GetHash64(const u8 *src, int len, u32 samples)
{
const u32 m = 0x5bd1e995;
const int r = 24;
u32 h1 = len;
u32 h2 = 0;
u32 Step = (len / 4);
const u32 * data = (const u32 *)src;
const u32 * end = data + Step;
const u8 * uEnd = (const u8 *)end;
if(samples == 0) samples = Step;
Step = Step / samples;
if(Step < 2) Step = 2;
while(data < end)
{
u32 k1 = data[0];
k1 *= m;
k1 ^= k1 >> r;
k1 *= m;
h1 *= m;
h1 ^= k1;
u32 k2 = data[1];
k2 *= m;
k2 ^= k2 >> r;
k2 *= m;
h2 *= m;
h2 ^= k2;
data+=Step;
}
if((len & 7) > 3)
{
u32 k1 = *(end - 1);
k1 *= m;
k1 ^= k1 >> r;
k1 *= m;
h1 *= m;
h1 ^= k1;
len -= 4;
}
switch(len & 3)
{
case 3: h2 ^= uEnd[2] << 16;
case 2: h2 ^= uEnd[1] << 8;
case 1: h2 ^= uEnd[0];
h2 *= m;
};
h1 ^= h2 >> 18; h1 *= m;
h2 ^= h1 >> 22; h2 *= m;
h1 ^= h2 >> 17; h1 *= m;
h2 ^= h1 >> 19; h2 *= m;
u64 h = h1;
h = (h << 32) | h2;
return h;
}
#endif
int TexDecoder_GetBlockWidthInTexels(u32 format)
{
switch (format)

5
Source/Core/VideoCommon/Src/TextureDecoder.h
View File

@ -17,7 +17,7 @@
#ifndef _TEXTUREDECODER_H
#define _TEXTUREDECODER_H
#include "hash.h"
enum
{
TMEM_SIZE = 1024*1024,
@ -86,9 +86,6 @@ enum PC_TexFormat
PC_TexFormat TexDecoder_Decode(u8 *dst, const u8 *src, int width, int height, int texformat, int tlutaddr, int tlutfmt,bool rgbaOnly = false);
PC_TexFormat GetPC_TexFormat(int texformat, int tlutfmt);
void TexDecoder_DecodeTexel(u8 *dst, const u8 *src, int s, int t, int imageWidth, int texformat, int tlutaddr, int tlutfmt);
u64 TexDecoder_GetHash64(const u8 *src, int len, u32 samples = 0);
void TexDecoder_SetTexFmtOverlayOptions(bool enable, bool center);
#endif

4
Source/Core/VideoCommon/Src/VertexShaderGen.cpp
View File

@ -45,10 +45,6 @@ void GetVertexShaderId(VERTEXSHADERUID *uid, u32 components)
(u32)xfregs.colChans[i].alpha.matsource) << 15;
}
// fog
uid->values[1] |= (((u32)bpmem.fog.c_proj_fsel.fsel & 3) << 30);
uid->values[2] |= (((u32)bpmem.fog.c_proj_fsel.fsel >> 2) << 30);
u32 *pcurvalue = &uid->values[3];
for (int i = 0; i < xfregs.numTexGens; ++i) {
TexMtxInfo tinfo = xfregs.texcoords[i].texmtxinfo;