- Speed up MSAA by using custom resolve shaders instead of ID3D11DeviceContext::ResolveSubresource().
- Fix EFB access when MSAA is enabled.
- Implement the "force linear filtering" option.
- Compile shaders using shader model 4.1 or 5.0 if available.
- Some minor cleanups.

Who cares about DX9 and OGL anyway :P

git-svn-id: https://dolphin-emu.googlecode.com/svn/trunk@6481 8ced0084-cf51-0410-be5f-012b33b47a6e
This commit is contained in:
NeoBrainX 2010-11-27 11:11:05 +00:00
parent 5340b4cb95
commit 7698d85a8e
11 changed files with 156 additions and 60 deletions

View File

@ -146,8 +146,6 @@ struct TargetRectangle : public MathUtil::Rectangle<int>
#define LOG_VTX()
bool IsD3D();
typedef enum
{
API_OPENGL,

View File

@ -235,7 +235,6 @@ HRESULT Create(HWND wnd)
aa_modes.clear();
EnumAAModes(adapter, aa_modes);
// this will need to be changed once multisampling gets implemented
DXGI_SWAP_CHAIN_DESC swap_chain_desc;
memset(&swap_chain_desc, 0, sizeof(swap_chain_desc));
swap_chain_desc.BufferCount = 1;
@ -337,8 +336,19 @@ void Close()
}
/* just returning the 4_0 ones here */
const char* VertexShaderVersionString() { return "vs_4_0"; }
const char* PixelShaderVersionString() { return "ps_4_0"; }
const char* VertexShaderVersionString()
{
if(featlevel == D3D_FEATURE_LEVEL_11_0) return "vs_5_0";
else if(featlevel == D3D_FEATURE_LEVEL_10_1) return "vs_4_1";
else /*if(featlevel == D3D_FEATURE_LEVEL_10_0)*/ return "vs_4_0";
}
const char* PixelShaderVersionString()
{
if(featlevel == D3D_FEATURE_LEVEL_11_0) return "ps_5_0";
else if(featlevel == D3D_FEATURE_LEVEL_10_1) return "ps_4_1";
else /*if(featlevel == D3D_FEATURE_LEVEL_10_0)*/ return "ps_4_0";
}
D3DTexture2D* &GetBackBuffer() { return backbuf; }
unsigned int GetBackBufferWidth() { return xres; }
@ -352,11 +362,11 @@ unsigned int GetMaxTextureSize()
switch (featlevel)
{
case D3D_FEATURE_LEVEL_11_0:
return 16384;
return D3D11_REQ_TEXTURE2D_U_OR_V_DIMENSION;
case D3D_FEATURE_LEVEL_10_1:
case D3D_FEATURE_LEVEL_10_0:
return 8192;
return D3D10_REQ_TEXTURE2D_U_OR_V_DIMENSION;
case D3D_FEATURE_LEVEL_9_3:
return 4096;

View File

@ -299,7 +299,6 @@ int CD3DFont::Init()
CHECK(hr==S_OK, "Create font blend state");
D3D::SetDebugObjectName((ID3D11DeviceChild*)m_blendstate, "blend state of a CD3DFont object");
// this might need to be changed when adding multisampling support
D3D11_RASTERIZER_DESC rastdesc = CD3D11_RASTERIZER_DESC(D3D11_FILL_SOLID, D3D11_CULL_NONE, false, 0, 0.f, 0.f, false, false, false, false);
hr = D3D::device->CreateRasterizerState(&rastdesc, &m_raststate);
CHECK(hr==S_OK, "Create font rasterizer state");
@ -470,12 +469,12 @@ void InitUtils()
util_vbuf = new UtilVertexBuffer(0x4000);
float border[4] = { 0.f, 0.f, 0.f, 0.f };
D3D11_SAMPLER_DESC samDesc = CD3D11_SAMPLER_DESC(D3D11_FILTER_MIN_MAG_MIP_POINT, D3D11_TEXTURE_ADDRESS_BORDER, D3D11_TEXTURE_ADDRESS_BORDER, D3D11_TEXTURE_ADDRESS_BORDER, 0.f, 1, D3D11_COMPARISON_ALWAYS, border, -D3D11_FLOAT32_MAX, D3D11_FLOAT32_MAX);
D3D11_SAMPLER_DESC samDesc = CD3D11_SAMPLER_DESC(D3D11_FILTER_MIN_MAG_MIP_POINT, D3D11_TEXTURE_ADDRESS_BORDER, D3D11_TEXTURE_ADDRESS_BORDER, D3D11_TEXTURE_ADDRESS_BORDER, 0.f, 1, D3D11_COMPARISON_ALWAYS, border, 0.f, 0.f);
HRESULT hr = D3D::device->CreateSamplerState(&samDesc, &point_copy_sampler);
if (FAILED(hr)) PanicAlert("Failed to create sampler state at %s %d\n", __FILE__, __LINE__);
else SetDebugObjectName((ID3D11DeviceChild*)point_copy_sampler, "point copy sampler state");
samDesc = CD3D11_SAMPLER_DESC(D3D11_FILTER_MIN_MAG_MIP_LINEAR, D3D11_TEXTURE_ADDRESS_BORDER, D3D11_TEXTURE_ADDRESS_BORDER, D3D11_TEXTURE_ADDRESS_BORDER, 0.f, 1, D3D11_COMPARISON_ALWAYS, border, -D3D11_FLOAT32_MAX, D3D11_FLOAT32_MAX);
samDesc = CD3D11_SAMPLER_DESC(D3D11_FILTER_MIN_MAG_MIP_LINEAR, D3D11_TEXTURE_ADDRESS_BORDER, D3D11_TEXTURE_ADDRESS_BORDER, D3D11_TEXTURE_ADDRESS_BORDER, 0.f, 1, D3D11_COMPARISON_ALWAYS, border, 0.f, 0.f);
hr = D3D::device->CreateSamplerState(&samDesc, &linear_copy_sampler);
if (FAILED(hr)) PanicAlert("Failed to create sampler state at %s %d\n", __FILE__, __LINE__);
else SetDebugObjectName((ID3D11DeviceChild*)linear_copy_sampler, "linear copy sampler state");

View File

@ -172,7 +172,7 @@ void XFBSource::Draw(const MathUtil::Rectangle<float> &sourcerc,
const MathUtil::Rectangle<float> &drawrc, int width, int height) const
{
D3D::drawShadedTexSubQuad(tex->GetSRV(), &sourcerc,
texWidth, texHeight, &drawrc, PixelShaderCache::GetColorCopyProgram(),
texWidth, texHeight, &drawrc, PixelShaderCache::GetColorCopyProgram(false),
VertexShaderCache::GetSimpleVertexShader(), VertexShaderCache::GetSimpleInputLayout());
}
@ -193,7 +193,7 @@ void XFBSource::CopyEFB()
D3D::drawShadedTexQuad(FramebufferManager::GetEFBColorTexture()->GetSRV(), sourceRc.AsRECT(),
Renderer::GetFullTargetWidth(), Renderer::GetFullTargetHeight(),
PixelShaderCache::GetColorCopyProgram(), VertexShaderCache::GetSimpleVertexShader(),
PixelShaderCache::GetColorCopyProgram(true), VertexShaderCache::GetSimpleVertexShader(),
VertexShaderCache::GetSimpleInputLayout());
D3D::context->OMSetRenderTargets(1, &FramebufferManager::GetEFBColorTexture()->GetRTV(),

View File

@ -40,9 +40,9 @@ const PixelShaderCache::PSCacheEntry* PixelShaderCache::last_entry;
LinearDiskCache<PIXELSHADERUID, u8> g_ps_disk_cache;
ID3D11PixelShader* s_ColorMatrixProgram = NULL;
ID3D11PixelShader* s_ColorCopyProgram = NULL;
ID3D11PixelShader* s_DepthMatrixProgram = NULL;
ID3D11PixelShader* s_ColorMatrixProgram[2] = {NULL};
ID3D11PixelShader* s_ColorCopyProgram[2] = {NULL};
ID3D11PixelShader* s_DepthMatrixProgram[2] = {NULL};
ID3D11PixelShader* s_ClearProgram = NULL;
const char clear_program_code[] = {
@ -65,6 +65,23 @@ const char color_copy_program_code[] = {
"}\n"
};
// TODO: Improve sampling algorithm!
const char color_copy_program_code_msaa[] = {
"sampler samp0 : register(s0);\n"
"Texture2DMS<float4, %d> Tex0 : register(t0);\n"
"void main(\n"
"out float4 ocol0 : SV_Target,\n"
"in float4 pos : SV_Position,\n"
"in float2 uv0 : TEXCOORD0){\n"
"int width, height, samples;\n"
"Tex0.GetDimensions(width, height, samples);\n"
"ocol0 = 0;\n"
"for(int i = 0; i < samples; ++i)\n"
" ocol0 += Tex0.Load(int2(uv0.x*(width), uv0.y*(height)), i);\n"
"ocol0 /= samples;\n"
"}\n"
};
const char color_matrix_program_code[] = {
"sampler samp0 : register(s0);\n"
"Texture2D Tex0 : register(t0);\n"
@ -78,6 +95,24 @@ const char color_matrix_program_code[] = {
"}\n"
};
const char color_matrix_program_code_msaa[] = {
"sampler samp0 : register(s0);\n"
"Texture2DMS<float4, %d> Tex0 : register(t0);\n"
"uniform float4 cColMatrix[5] : register(c0);\n"
"void main(\n"
"out float4 ocol0 : SV_Target,\n"
"in float4 pos : SV_Position,\n"
" in float2 uv0 : TEXCOORD0){\n"
"int width, height, samples;\n"
"Tex0.GetDimensions(width, height, samples);\n"
"float4 texcol = 0;\n"
"for(int i = 0; i < samples; ++i)\n"
" texcol += Tex0.Load(int2(uv0.x*(width), uv0.y*(height)), i);\n"
"texcol /= samples;\n"
"ocol0 = float4(dot(texcol,cColMatrix[0]),dot(texcol,cColMatrix[1]),dot(texcol,cColMatrix[2]),dot(texcol,cColMatrix[3])) + cColMatrix[4];\n"
"}\n"
};
const char depth_matrix_program[] = {
"sampler samp0 : register(s0);\n"
"Texture2D Tex0 : register(t0);\n"
@ -93,19 +128,69 @@ const char depth_matrix_program[] = {
"}\n"
};
ID3D11PixelShader* PixelShaderCache::GetColorMatrixProgram()
const char depth_matrix_program_msaa[] = {
"sampler samp0 : register(s0);\n"
"Texture2DMS<float4, %d> Tex0 : register(t0);\n"
"uniform float4 cColMatrix[5] : register(c0);\n"
"void main(\n"
"out float4 ocol0 : SV_Target,\n"
" in float4 pos : SV_Position,\n"
" in float2 uv0 : TEXCOORD0){\n"
"int width, height, samples;\n"
"Tex0.GetDimensions(width, height, samples);\n"
"float4 texcol = 0;\n"
"for(int i = 0; i < samples; ++i)\n"
" texcol += Tex0.Load(int2(uv0.x*(width), uv0.y*(height)), i);\n"
"texcol /= samples;\n"
"float4 EncodedDepth = frac((texcol.r * (16777215.0f/16777216.0f)) * float4(1.0f,255.0f,255.0f*255.0f,255.0f*255.0f*255.0f));\n"
"texcol = float4((EncodedDepth.rgb * (16777216.0f/16777215.0f)),1.0f);\n"
"ocol0 = float4(dot(texcol,cColMatrix[0]),dot(texcol,cColMatrix[1]),dot(texcol,cColMatrix[2]),dot(texcol,cColMatrix[3])) + cColMatrix[4];\n"
"}\n"
};
ID3D11PixelShader* PixelShaderCache::GetColorCopyProgram(bool multisampled)
{
return s_ColorMatrixProgram;
if (!multisampled || s_ColorCopyProgram[1]) return s_ColorCopyProgram[multisampled];
else
{
// create MSAA shader for current AA mode
char buf[1024];
int l = sprintf_s(buf, 1024, color_copy_program_code_msaa, D3D::GetAAMode(g_ActiveConfig.iMultisampleMode));
s_ColorCopyProgram[1] = D3D::CompileAndCreatePixelShader(buf, l);
CHECK(s_ColorCopyProgram[1]!=NULL, "Create color copy MSAA pixel shader");
D3D::SetDebugObjectName((ID3D11DeviceChild*)s_ColorCopyProgram[1], "color copy MSAA pixel shader");
return s_ColorCopyProgram[1];
}
}
ID3D11PixelShader* PixelShaderCache::GetDepthMatrixProgram()
ID3D11PixelShader* PixelShaderCache::GetColorMatrixProgram(bool multisampled)
{
return s_DepthMatrixProgram;
if (!multisampled || s_ColorMatrixProgram[1]) return s_ColorMatrixProgram[multisampled];
else
{
// create MSAA shader for current AA mode
char buf[1024];
int l = sprintf_s(buf, 1024, color_matrix_program_code_msaa, D3D::GetAAMode(g_ActiveConfig.iMultisampleMode));
s_ColorMatrixProgram[1] = D3D::CompileAndCreatePixelShader(buf, l);
CHECK(s_ColorMatrixProgram[1]!=NULL, "Create color matrix MSAA pixel shader");
D3D::SetDebugObjectName((ID3D11DeviceChild*)s_ColorMatrixProgram[1], "color matrix MSAA pixel shader");
return s_ColorMatrixProgram[1];
}
}
ID3D11PixelShader* PixelShaderCache::GetColorCopyProgram()
ID3D11PixelShader* PixelShaderCache::GetDepthMatrixProgram(bool multisampled)
{
return s_ColorCopyProgram;
if (!multisampled || s_DepthMatrixProgram[1]) return s_DepthMatrixProgram[multisampled];
else
{
// create MSAA shader for current AA mode
char buf[1024];
int l = sprintf_s(buf, 1024, depth_matrix_program_msaa, D3D::GetAAMode(g_ActiveConfig.iMultisampleMode));
s_DepthMatrixProgram[1] = D3D::CompileAndCreatePixelShader(buf, l);
CHECK(s_DepthMatrixProgram[1]!=NULL, "Create depth matrix MSAA pixel shader");
D3D::SetDebugObjectName((ID3D11DeviceChild*)s_DepthMatrixProgram[1], "depth matrix MSAA pixel shader");
return s_DepthMatrixProgram[1];
}
}
ID3D11PixelShader* PixelShaderCache::GetClearProgram()
@ -175,19 +260,19 @@ void PixelShaderCache::Init()
D3D::SetDebugObjectName((ID3D11DeviceChild*)s_ClearProgram, "clear pixel shader");
// used when copying/resolving the color buffer
s_ColorCopyProgram = D3D::CompileAndCreatePixelShader(color_copy_program_code, sizeof(color_copy_program_code));
CHECK(s_ColorCopyProgram!=NULL, "Create color copy pixel shader");
D3D::SetDebugObjectName((ID3D11DeviceChild*)s_ClearProgram, "color copy pixel shader");
s_ColorCopyProgram[0] = D3D::CompileAndCreatePixelShader(color_copy_program_code, sizeof(color_copy_program_code));
CHECK(s_ColorCopyProgram[0]!=NULL, "Create color copy pixel shader");
D3D::SetDebugObjectName((ID3D11DeviceChild*)s_ColorCopyProgram[0], "color copy pixel shader");
// used for color conversion
s_ColorMatrixProgram = D3D::CompileAndCreatePixelShader(color_matrix_program_code, sizeof(color_matrix_program_code));
CHECK(s_ColorMatrixProgram!=NULL, "Create color matrix pixel shader");
D3D::SetDebugObjectName((ID3D11DeviceChild*)s_ClearProgram, "color matrix pixel shader");
s_ColorMatrixProgram[0] = D3D::CompileAndCreatePixelShader(color_matrix_program_code, sizeof(color_matrix_program_code));
CHECK(s_ColorMatrixProgram[0]!=NULL, "Create color matrix pixel shader");
D3D::SetDebugObjectName((ID3D11DeviceChild*)s_ColorMatrixProgram[0], "color matrix pixel shader");
// used for depth copy
s_DepthMatrixProgram = D3D::CompileAndCreatePixelShader(depth_matrix_program, sizeof(depth_matrix_program));
CHECK(s_DepthMatrixProgram!=NULL, "Create depth matrix pixel shader");
D3D::SetDebugObjectName((ID3D11DeviceChild*)s_ClearProgram, "depth matrix pixel shader");
s_DepthMatrixProgram[0] = D3D::CompileAndCreatePixelShader(depth_matrix_program, sizeof(depth_matrix_program));
CHECK(s_DepthMatrixProgram[0]!=NULL, "Create depth matrix pixel shader");
D3D::SetDebugObjectName((ID3D11DeviceChild*)s_DepthMatrixProgram[0], "depth matrix pixel shader");
Clear();
@ -211,12 +296,23 @@ void PixelShaderCache::Clear()
PixelShaders.clear();
}
// Used in Swap() when AA mode has changed
void PixelShaderCache::InvalidateMSAAShaders()
{
SAFE_RELEASE(s_ColorCopyProgram[1]);
SAFE_RELEASE(s_ColorMatrixProgram[1]);
SAFE_RELEASE(s_DepthMatrixProgram[1]);
}
void PixelShaderCache::Shutdown()
{
SAFE_RELEASE(s_ColorMatrixProgram);
SAFE_RELEASE(s_ColorCopyProgram);
SAFE_RELEASE(s_DepthMatrixProgram);
SAFE_RELEASE(s_ClearProgram);
for (int i = 0; i < 2; ++i)
{
SAFE_RELEASE(s_ColorCopyProgram[i]);
SAFE_RELEASE(s_ColorMatrixProgram[i]);
SAFE_RELEASE(s_DepthMatrixProgram[i]);
}
Clear();
g_ps_disk_cache.Sync();

View File

@ -35,11 +35,13 @@ public:
static bool SetShader(DSTALPHA_MODE dstAlphaMode, u32 components);
static bool InsertByteCode(const PIXELSHADERUID &uid, const void* bytecode, unsigned int bytecodelen);
static ID3D11PixelShader* GetColorMatrixProgram();
static ID3D11PixelShader* GetColorCopyProgram();
static ID3D11PixelShader* GetDepthMatrixProgram();
static ID3D11PixelShader* GetColorMatrixProgram(bool multisampled);
static ID3D11PixelShader* GetColorCopyProgram(bool multisampled);
static ID3D11PixelShader* GetDepthMatrixProgram(bool multisampled);
static ID3D11PixelShader* GetClearProgram();
static void InvalidateMSAAShaders();
private:
struct PSCacheEntry
{

View File

@ -283,7 +283,6 @@ void SetupDeviceObjects()
CHECK(hr==S_OK, "Create depth state for Renderer::ResetAPIState");
D3D::SetDebugObjectName((ID3D11DeviceChild*)resetdepthstate, "depth stencil state for Renderer::ResetAPIState");
// this might need to be changed once multisampling support gets added
D3D11_RASTERIZER_DESC rastdesc = CD3D11_RASTERIZER_DESC(D3D11_FILL_SOLID, D3D11_CULL_NONE, false, 0, 0.f, 0.f, false, false, false, false);
hr = D3D::device->CreateRasterizerState(&rastdesc, &resetraststate);
CHECK(hr==S_OK, "Create rasterizer state for Renderer::ResetAPIState");
@ -546,11 +545,11 @@ u32 Renderer::AccessEFB(EFBAccessType type, u32 x, u32 y, u32 poke_data)
D3D::context->PSSetConstantBuffers(0, 1, &access_efb_cbuf);
D3D::context->OMSetRenderTargets(1, &FramebufferManager::GetEFBDepthReadTexture()->GetRTV(), NULL);
D3D::SetPointCopySampler();
D3D::drawShadedTexQuad(FramebufferManager::GetResolvedEFBDepthTexture()->GetSRV(),
D3D::drawShadedTexQuad(FramebufferManager::GetEFBDepthTexture()->GetSRV(),
&RectToLock,
Renderer::GetFullTargetWidth(),
Renderer::GetFullTargetHeight(),
PixelShaderCache::GetDepthMatrixProgram(),
PixelShaderCache::GetDepthMatrixProgram(true),
VertexShaderCache::GetSimpleVertexShader(),
VertexShaderCache::GetSimpleInputLayout());
@ -883,8 +882,10 @@ void Renderer::Swap(u32 xfbAddr, FieldType field, u32 fbWidth, u32 fbHeight,cons
else
{
TargetRectangle targetRc = Renderer::ConvertEFBRectangle(rc);
// TODO: Improve sampling algorithm for the pixel shader so that we can use the multisampled EFB texture as source
D3DTexture2D* read_texture = FramebufferManager::GetResolvedEFBColorTexture();
D3D::drawShadedTexQuad(read_texture->GetSRV(), targetRc.AsRECT(), Renderer::GetFullTargetWidth(), Renderer::GetFullTargetHeight(), PixelShaderCache::GetColorCopyProgram(),VertexShaderCache::GetSimpleVertexShader(), VertexShaderCache::GetSimpleInputLayout());
D3D::drawShadedTexQuad(read_texture->GetSRV(), targetRc.AsRECT(), Renderer::GetFullTargetWidth(), Renderer::GetFullTargetHeight(), PixelShaderCache::GetColorCopyProgram(false),VertexShaderCache::GetSimpleVertexShader(), VertexShaderCache::GetSimpleInputLayout());
}
// done with drawing the game stuff, good moment to save a screenshot
if (s_bScreenshot)
@ -964,6 +965,7 @@ void Renderer::Swap(u32 xfbAddr, FieldType field, u32 fbWidth, u32 fbHeight,cons
s_LastAA != g_ActiveConfig.iMultisampleMode)
{
s_LastAA = g_ActiveConfig.iMultisampleMode;
PixelShaderCache::InvalidateMSAAShaders();
// TODO: Aren't we still holding a reference to the back buffer right now?
D3D::Reset();
@ -1079,7 +1081,11 @@ void Renderer::SetSamplerState(int stage, int texindex)
// TODO: Clarify whether these values are correct
// NOTE: since there's no "no filter" in DX11 we're using point filters in these cases
if (tm0.min_filter & 4) // linear min filter
if (g_ActiveConfig.bForceFiltering)
{
D3D::gfxstate->SetSamplerFilter(stage, D3D11_FILTER_MIN_MAG_MIP_LINEAR);
}
else if (tm0.min_filter & 4) // linear min filter
{
if (tm0.mag_filter) // linear mag filter
{

View File

@ -141,9 +141,9 @@ void TextureCache::TCacheEntry::FromRenderTarget(bool bFromZBuffer, bool bScaleB
D3D::context->OMSetRenderTargets(1, &texture->GetRTV(), NULL);
D3D::drawShadedTexQuad(
(bFromZBuffer) ? FramebufferManager::GetResolvedEFBDepthTexture()->GetSRV() : FramebufferManager::GetResolvedEFBColorTexture()->GetSRV(),
(bFromZBuffer) ? FramebufferManager::GetEFBDepthTexture()->GetSRV() : FramebufferManager::GetEFBColorTexture()->GetSRV(),
&sourcerect, Renderer::GetFullTargetWidth(), Renderer::GetFullTargetHeight(),
(bFromZBuffer) ? PixelShaderCache::GetDepthMatrixProgram() : PixelShaderCache::GetColorMatrixProgram(),
(bFromZBuffer) ? PixelShaderCache::GetDepthMatrixProgram(true) : PixelShaderCache::GetColorMatrixProgram(true),
VertexShaderCache::GetSimpleVertexShader(), VertexShaderCache::GetSimpleInputLayout());
D3D::context->OMSetRenderTargets(1, &FramebufferManager::GetEFBColorTexture()->GetRTV(), FramebufferManager::GetEFBDepthTexture()->GetDSV());

View File

@ -57,11 +57,6 @@
HINSTANCE g_hInstance = NULL;
bool IsD3D()
{
return true;
}
// This is used for the functions right below here which use wxwidgets
#if defined(HAVE_WX) && HAVE_WX
WXDLLIMPEXP_BASE void wxSetInstance(HINSTANCE hInst);

View File

@ -56,11 +56,6 @@
HINSTANCE g_hInstance = NULL;
bool IsD3D()
{
return true;
}
// This is used for the functions right below here which use wxwidgets
#if defined(HAVE_WX) && HAVE_WX
WXDLLIMPEXP_BASE void wxSetInstance(HINSTANCE hInst);

View File

@ -101,11 +101,6 @@ int GLScissorX, GLScissorY, GLScissorW, GLScissorH;
static volatile u32 s_doStateRequested = FALSE;
#endif
bool IsD3D()
{
return false;
}
// This is used for the functions right below here which use wxwidgets
#if defined(HAVE_WX) && HAVE_WX
#ifdef _WIN32