dolphin/Source/Plugins/Plugin_VideoOGL/Src/VertexManager.cpp
rodolfoosvaldobogado@gmail.com 5230146c73 Hey, long time no commits :).
So to compensate lets bring back some speed to the emulation.
change a little the way the vertex are send to the gpu,
This first implementation changes dx9 a lot and dx11 a little to increase the parallelism between the cpu and gpu.
ogl: is my next step in ogl is a little more trickier so i have to take a little more time.
the original concept is Marcos idea, with my little touch to make it even more faster.
what to look for: SPEEEEEDDD :).
please test it a lot and let me know if you see any problem.
in dx9 the code is prepared to fall back to the previous implementation if your card does not support the amount of buffers needed.
So if you did not experience any speed gains you know where is the problem :).
for the ones with more experience and compression of the code please test changing the amount and size of the buffers to tune this for your specific machine.
The current values are the sweet spot for my machine.
All must Thanks Marcos, I hate him for giving good ideas when I'm full of work.
2012-10-20 10:22:15 -03:00

278 lines
8.7 KiB
C++

// Copyright (C) 2003 Dolphin Project.
// 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, version 2.0.
// 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 2.0 for more details.
// A copy of the GPL 2.0 should have been included with the program.
// If not, see http://www.gnu.org/licenses/
// Official SVN repository and contact information can be found at
// http://code.google.com/p/dolphin-emu/
#include "Globals.h"
#include <fstream>
#include <vector>
#include "Fifo.h"
#include "VideoConfig.h"
#include "Statistics.h"
#include "MemoryUtil.h"
#include "Render.h"
#include "ImageWrite.h"
#include "BPMemory.h"
#include "TextureCache.h"
#include "PixelShaderCache.h"
#include "PixelShaderManager.h"
#include "VertexShaderCache.h"
#include "VertexShaderManager.h"
#include "VertexShaderGen.h"
#include "VertexLoader.h"
#include "VertexManager.h"
#include "IndexGenerator.h"
#include "OpcodeDecoding.h"
#include "FileUtil.h"
#include "Debugger.h"
#include "main.h"
// internal state for loading vertices
extern NativeVertexFormat *g_nativeVertexFmt;
namespace OGL
{
//static GLint max_Index_size = 0;
//static GLuint s_vboBuffers[MAXVBOBUFFERCOUNT] = {0};
//static int s_nCurVBOIndex = 0; // current free buffer
VertexManager::VertexManager()
{
// TODO: doesn't seem to be used anywhere
//glGetIntegerv(GL_MAX_ELEMENTS_INDICES, (GLint*)&max_Index_size);
//
//if (max_Index_size > MAXIBUFFERSIZE)
// max_Index_size = MAXIBUFFERSIZE;
//
//GL_REPORT_ERRORD();
glEnableClientState(GL_VERTEX_ARRAY);
GL_REPORT_ERRORD();
}
void VertexManager::CreateDeviceObjects()
{
}
void VertexManager::DestroyDeviceObjects()
{
}
void VertexManager::Draw()
{
if (IndexGenerator::GetNumTriangles() > 0)
{
glDrawElements(GL_TRIANGLES, IndexGenerator::GetTriangleindexLen(), GL_UNSIGNED_SHORT, TIBuffer);
INCSTAT(stats.thisFrame.numIndexedDrawCalls);
}
if (IndexGenerator::GetNumLines() > 0)
{
glDrawElements(GL_LINES, IndexGenerator::GetLineindexLen(), GL_UNSIGNED_SHORT, LIBuffer);
INCSTAT(stats.thisFrame.numIndexedDrawCalls);
}
if (IndexGenerator::GetNumPoints() > 0)
{
glDrawElements(GL_POINTS, IndexGenerator::GetPointindexLen(), GL_UNSIGNED_SHORT, PIBuffer);
INCSTAT(stats.thisFrame.numIndexedDrawCalls);
}
}
void VertexManager::vFlush()
{
if (LocalVBuffer == s_pCurBufferPointer) return;
if (Flushed) return;
Flushed=true;
VideoFifo_CheckEFBAccess();
#if defined(_DEBUG) || defined(DEBUGFAST)
PRIM_LOG("frame%d:\n texgen=%d, numchan=%d, dualtex=%d, ztex=%d, cole=%d, alpe=%d, ze=%d", g_ActiveConfig.iSaveTargetId, xfregs.numTexGen.numTexGens,
xfregs.numChan.numColorChans, xfregs.dualTexTrans.enabled, bpmem.ztex2.op,
bpmem.blendmode.colorupdate, bpmem.blendmode.alphaupdate, bpmem.zmode.updateenable);
for (unsigned int i = 0; i < xfregs.numChan.numColorChans; ++i)
{
LitChannel* ch = &xfregs.color[i];
PRIM_LOG("colchan%d: matsrc=%d, light=0x%x, ambsrc=%d, diffunc=%d, attfunc=%d", i, ch->matsource, ch->GetFullLightMask(), ch->ambsource, ch->diffusefunc, ch->attnfunc);
ch = &xfregs.alpha[i];
PRIM_LOG("alpchan%d: matsrc=%d, light=0x%x, ambsrc=%d, diffunc=%d, attfunc=%d", i, ch->matsource, ch->GetFullLightMask(), ch->ambsource, ch->diffusefunc, ch->attnfunc);
}
for (unsigned int i = 0; i < xfregs.numTexGen.numTexGens; ++i)
{
TexMtxInfo tinfo = xfregs.texMtxInfo[i];
if (tinfo.texgentype != XF_TEXGEN_EMBOSS_MAP) tinfo.hex &= 0x7ff;
if (tinfo.texgentype != XF_TEXGEN_REGULAR) tinfo.projection = 0;
PRIM_LOG("txgen%d: proj=%d, input=%d, gentype=%d, srcrow=%d, embsrc=%d, emblght=%d, postmtx=%d, postnorm=%d",
i, tinfo.projection, tinfo.inputform, tinfo.texgentype, tinfo.sourcerow, tinfo.embosssourceshift, tinfo.embosslightshift,
xfregs.postMtxInfo[i].index, xfregs.postMtxInfo[i].normalize);
}
PRIM_LOG("pixel: tev=%d, ind=%d, texgen=%d, dstalpha=%d, alphafunc=0x%x", bpmem.genMode.numtevstages+1, bpmem.genMode.numindstages,
bpmem.genMode.numtexgens, (u32)bpmem.dstalpha.enable, (bpmem.alphaFunc.hex>>16)&0xff);
#endif
(void)GL_REPORT_ERROR();
//glBindBuffer(GL_ARRAY_BUFFER, s_vboBuffers[s_nCurVBOIndex]);
//glBufferData(GL_ARRAY_BUFFER, s_pCurBufferPointer - LocalVBuffer, LocalVBuffer, GL_STREAM_DRAW);
GL_REPORT_ERRORD();
// setup the pointers
if (g_nativeVertexFmt)
g_nativeVertexFmt->SetupVertexPointers();
GL_REPORT_ERRORD();
u32 usedtextures = 0;
for (u32 i = 0; i < (u32)bpmem.genMode.numtevstages + 1; ++i)
if (bpmem.tevorders[i / 2].getEnable(i & 1))
usedtextures |= 1 << bpmem.tevorders[i/2].getTexMap(i & 1);
if (bpmem.genMode.numindstages > 0)
for (u32 i = 0; i < (u32)bpmem.genMode.numtevstages + 1; ++i)
if (bpmem.tevind[i].IsActive() && bpmem.tevind[i].bt < bpmem.genMode.numindstages)
usedtextures |= 1 << bpmem.tevindref.getTexMap(bpmem.tevind[i].bt);
for (int i = 0; i < 8; i++)
{
if (usedtextures & (1 << i))
{
glActiveTexture(GL_TEXTURE0 + i);
FourTexUnits &tex = bpmem.tex[i >> 2];
TextureCache::TCacheEntryBase* tentry = TextureCache::Load(i,
(tex.texImage3[i&3].image_base/* & 0x1FFFFF*/) << 5,
tex.texImage0[i&3].width + 1, tex.texImage0[i&3].height + 1,
tex.texImage0[i&3].format, tex.texTlut[i&3].tmem_offset<<9,
tex.texTlut[i&3].tlut_format,
(tex.texMode0[i&3].min_filter & 3) && (tex.texMode0[i&3].min_filter != 8),
tex.texMode1[i&3].max_lod >> 4,
tex.texImage1[i&3].image_type);
if (tentry)
{
// 0s are probably for no manual wrapping needed.
PixelShaderManager::SetTexDims(i, tentry->native_width, tentry->native_height, 0, 0);
}
else
ERROR_LOG(VIDEO, "error loading texture");
}
}
// set global constants
VertexShaderManager::SetConstants();
PixelShaderManager::SetConstants();
bool useDstAlpha = !g_ActiveConfig.bDstAlphaPass && bpmem.dstalpha.enable && bpmem.blendmode.alphaupdate
&& bpmem.zcontrol.pixel_format == PIXELFMT_RGBA6_Z24;
#ifdef USE_DUAL_SOURCE_BLEND
bool dualSourcePossible = GLEW_ARB_blend_func_extended;
// finally bind
FRAGMENTSHADER* ps;
if (dualSourcePossible)
{
if (useDstAlpha)
{
// If host supports GL_ARB_blend_func_extended, we can do dst alpha in
// the same pass as regular rendering.
g_renderer->SetBlendMode(true);
ps = PixelShaderCache::SetShader(DSTALPHA_DUAL_SOURCE_BLEND, g_nativeVertexFmt->m_components);
}
else
{
g_renderer->SetBlendMode(true);
ps = PixelShaderCache::SetShader(DSTALPHA_NONE,g_nativeVertexFmt->m_components);
}
}
else
{
ps = PixelShaderCache::SetShader(DSTALPHA_NONE,g_nativeVertexFmt->m_components);
}
#else
bool dualSourcePossible = false;
FRAGMENTSHADER* ps = PixelShaderCache::SetShader(DSTALPHA_NONE,g_nativeVertexFmt->m_components);
#endif
VERTEXSHADER* vs = VertexShaderCache::SetShader(g_nativeVertexFmt->m_components);
if (ps) PixelShaderCache::SetCurrentShader(ps->glprogid); // Lego Star Wars crashes here.
if (vs) VertexShaderCache::SetCurrentShader(vs->glprogid);
Draw();
// run through vertex groups again to set alpha
if (useDstAlpha && !dualSourcePossible)
{
ps = PixelShaderCache::SetShader(DSTALPHA_ALPHA_PASS,g_nativeVertexFmt->m_components);
if (ps) PixelShaderCache::SetCurrentShader(ps->glprogid);
// only update alpha
glColorMask(GL_FALSE, GL_FALSE, GL_FALSE, GL_TRUE);
glDisable(GL_BLEND);
Draw();
// restore color mask
g_renderer->SetColorMask();
if (bpmem.blendmode.blendenable || bpmem.blendmode.subtract)
glEnable(GL_BLEND);
}
GFX_DEBUGGER_PAUSE_AT(NEXT_FLUSH, true);
//s_nCurVBOIndex = (s_nCurVBOIndex + 1) % ARRAYSIZE(s_vboBuffers);
s_pCurBufferPointer = LocalVBuffer;
IndexGenerator::Start(TIBuffer,LIBuffer,PIBuffer);
#if defined(_DEBUG) || defined(DEBUGFAST)
if (g_ActiveConfig.iLog & CONF_SAVESHADERS)
{
// save the shaders
char strfile[255];
sprintf(strfile, "%sps%.3d.txt", File::GetUserPath(D_DUMPFRAMES_IDX).c_str(), g_ActiveConfig.iSaveTargetId);
std::ofstream fps(strfile);
fps << ps->strprog.c_str();
sprintf(strfile, "%svs%.3d.txt", File::GetUserPath(D_DUMPFRAMES_IDX).c_str(), g_ActiveConfig.iSaveTargetId);
std::ofstream fvs(strfile);
fvs << vs->strprog.c_str();
}
if (g_ActiveConfig.iLog & CONF_SAVETARGETS)
{
char str[128];
sprintf(str, "%starg%.3d.tga", File::GetUserPath(D_DUMPFRAMES_IDX).c_str(), g_ActiveConfig.iSaveTargetId);
TargetRectangle tr;
tr.left = 0;
tr.right = Renderer::GetTargetWidth();
tr.top = 0;
tr.bottom = Renderer::GetTargetHeight();
g_renderer->SaveScreenshot(str, tr);
}
#endif
g_Config.iSaveTargetId++;
ClearEFBCache();
GL_REPORT_ERRORD();
}
} // namespace