Mirrors/yuzu
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions

shader/other: Implement thread comparisons (NV_shader_thread_group)

Browse source 
Hardware S2R special registers match gl_Thread*MaskNV. We can trivially
implement these using Nvidia's extension on OpenGL or naively stubbing
them with the ARB instructions to match. This might cause issues if the
host device warp size doesn't match Nvidia's. That said, this is
unlikely on proper shaders.

Refer to the attached url for more documentation about these flags.
https://www.khronos.org/registry/OpenGL/extensions/NV/NV_shader_thread_group.txt
This commit is contained in:
ReinUsesLisp 2020-05-15 01:43:44 -03:00
parent cf4ee279c6
commit e2b67a868b
4 changed files with 72 additions and 0 deletions
Download patch file Download diff file Expand all files Collapse all files

23
src/video_core/renderer_vulkan/vk_shader_decompiler.cpp
View file

@ -515,6 +515,16 @@ private:
void DeclareCommon() {
thread_id =
DeclareInputBuiltIn(spv::BuiltIn::SubgroupLocalInvocationId, t_in_uint, "thread_id");
thread_masks[0] =
DeclareInputBuiltIn(spv::BuiltIn::SubgroupEqMask, t_in_uint4, "thread_eq_mask");
thread_masks[1] =
DeclareInputBuiltIn(spv::BuiltIn::SubgroupGeMask, t_in_uint4, "thread_ge_mask");
thread_masks[2] =
DeclareInputBuiltIn(spv::BuiltIn::SubgroupGtMask, t_in_uint4, "thread_gt_mask");
thread_masks[3] =
DeclareInputBuiltIn(spv::BuiltIn::SubgroupLeMask, t_in_uint4, "thread_le_mask");
thread_masks[4] =
DeclareInputBuiltIn(spv::BuiltIn::SubgroupLtMask, t_in_uint4, "thread_lt_mask");
}
void DeclareVertex() {
@ -2175,6 +2185,13 @@ private:
return {OpLoad(t_uint, thread_id), Type::Uint};
}
template <std::size_t index>
Expression ThreadMask(Operation) {
// TODO(Rodrigo): Handle devices with different warp sizes
const Id mask = thread_masks[index];
return {OpLoad(t_uint, AccessElement(t_in_uint, mask, 0)), Type::Uint};
}
Expression ShuffleIndexed(Operation operation) {
const Id value = AsFloat(Visit(operation[0]));
const Id index = AsUint(Visit(operation[1]));
@ -2639,6 +2656,11 @@ private:
&SPIRVDecompiler::Vote<&Module::OpSubgroupAllEqualKHR>,
&SPIRVDecompiler::ThreadId,
&SPIRVDecompiler::ThreadMask<0>, // Eq
&SPIRVDecompiler::ThreadMask<1>, // Ge
&SPIRVDecompiler::ThreadMask<2>, // Gt
&SPIRVDecompiler::ThreadMask<3>, // Le
&SPIRVDecompiler::ThreadMask<4>, // Lt
&SPIRVDecompiler::ShuffleIndexed,
&SPIRVDecompiler::MemoryBarrierGL,
@ -2763,6 +2785,7 @@ private:
Id workgroup_id{};
Id local_invocation_id{};
Id thread_id{};
std::array<Id, 5> thread_masks{}; // eq, ge, gt, le, lt
VertexIndices in_indices;
VertexIndices out_indices;