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

PixelShaderGen: Split bbox into seperate variables 

The Metal shader compiler fails to compile the atomic instructions
when operating on individual components of a vector. Spltting it
into four variables shouldn't make any difference for other
platforms, as they are accessed independently.
This commit is contained in:
Stenzek 2018-08-29 03:00:08 -07:00
parent 673f1963a0
commit 0c0d66809d
2 changed files with 33 additions and 19 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

24
Source/Core/VideoCommon/PixelShaderGen.cpp
View File

@ -448,7 +448,7 @@ void WritePixelShaderCommonHeader(ShaderCode& out, APIType ApiType, u32 num_texg
if (ApiType == APIType::OpenGL || ApiType == APIType::Vulkan)
{
out.Write("SSBO_BINDING(0) buffer BBox {\n"
"\tint4 bbox_data;\n"
"\tint bbox_left, bbox_right, bbox_top, bbox_bottom;\n"
"};\n");
}
else
@ -853,13 +853,21 @@ ShaderCode GeneratePixelShaderCode(APIType ApiType, const ShaderHostConfig& host
if (uid_data->bounding_box)
{
const char* atomic_op =
(ApiType == APIType::OpenGL || ApiType == APIType::Vulkan) ? "atomic" : "Interlocked";
out.Write("\tif(bbox_data[0] > int(rawpos.x)) %sMin(bbox_data[0], int(rawpos.x));\n"
"\tif(bbox_data[1] < int(rawpos.x)) %sMax(bbox_data[1], int(rawpos.x));\n"
"\tif(bbox_data[2] > int(rawpos.y)) %sMin(bbox_data[2], int(rawpos.y));\n"
"\tif(bbox_data[3] < int(rawpos.y)) %sMax(bbox_data[3], int(rawpos.y));\n",
atomic_op, atomic_op, atomic_op, atomic_op);
if (ApiType == APIType::D3D)
{
out.Write(
"\tif(bbox_data[0] > int(rawpos.x)) InterlockedMin(bbox_data[0], int(rawpos.x));\n"
"\tif(bbox_data[1] < int(rawpos.x)) InterlockedMax(bbox_data[1], int(rawpos.x));\n"
"\tif(bbox_data[2] > int(rawpos.y)) InterlockedMin(bbox_data[2], int(rawpos.y));\n"
"\tif(bbox_data[3] < int(rawpos.y)) InterlockedMax(bbox_data[3], int(rawpos.y));\n");
}
else
{
out.Write("\tif(bbox_left > int(rawpos.x)) atomicMin(bbox_left, int(rawpos.x));\n"
"\tif(bbox_right < int(rawpos.x)) atomicMax(bbox_right, int(rawpos.x));\n"
"\tif(bbox_top > int(rawpos.y)) atomicMin(bbox_top, int(rawpos.y));\n"
"\tif(bbox_bottom < int(rawpos.y)) atomicMax(bbox_bottom, int(rawpos.y));\n");
}
}
out.Write("}\n");

28
Source/Core/VideoCommon/UberShaderPixel.cpp
View File

@ -1239,17 +1239,23 @@ ShaderCode GenPixelShader(APIType ApiType, const ShaderHostConfig& host_config,
if (bounding_box)
{
const char* atomic_op =
(ApiType == APIType::OpenGL || ApiType == APIType::Vulkan) ? "atomic" : "Interlocked";
out.Write(" if (bpmem_bounding_box) {\n");
out.Write(" if(bbox_data[0] > int(rawpos.x)) %sMin(bbox_data[0], int(rawpos.x));\n",
atomic_op);
out.Write(" if(bbox_data[1] < int(rawpos.x)) %sMax(bbox_data[1], int(rawpos.x));\n",
atomic_op);
out.Write(" if(bbox_data[2] > int(rawpos.y)) %sMin(bbox_data[2], int(rawpos.y));\n",
atomic_op);
out.Write(" if(bbox_data[3] < int(rawpos.y)) %sMax(bbox_data[3], int(rawpos.y));\n",
atomic_op);
if (ApiType == APIType::D3D)
{
out.Write(
" if(bbox_data[0] > int(rawpos.x)) InterlockedMin(bbox_data[0], int(rawpos.x));\n"
" if(bbox_data[1] < int(rawpos.x)) InterlockedMax(bbox_data[1], int(rawpos.x));\n"
" if(bbox_data[2] > int(rawpos.y)) InterlockedMin(bbox_data[2], int(rawpos.y));\n"
" if(bbox_data[3] < int(rawpos.y)) InterlockedMax(bbox_data[3], int(rawpos.y));\n");
}
else
{
out.Write("\tif(bbox_left > int(rawpos.x)) atomicMin(bbox_left, int(rawpos.x));\n"
"\tif(bbox_right < int(rawpos.x)) atomicMax(bbox_right, int(rawpos.x));\n"
"\tif(bbox_top > int(rawpos.y)) atomicMin(bbox_top, int(rawpos.y));\n"
"\tif(bbox_bottom < int(rawpos.y)) atomicMax(bbox_bottom, int(rawpos.y));\n");
}
out.Write(" }\n");
}
@ -1414,4 +1420,4 @@ void EnumeratePixelShaderUids(const std::function<void(const PixelShaderUid&)>&
}
}
}
}
} // namespace UberShader