[CK-Tile] move out memory operation from cshuffle epilogue class (#3359)

* initial poc

* factor out common parts in operator()

* cv4

* rest of the universal gemm pipelines

* fix test

* remove boilerplate from tile engine

* fix example

* fix example

* format

* fix tests build for gemm

* remove base pipeline codegen from gemm instance builder

* unify v3 logic with the rest of universal gemm pipelines

* fix build for multi abd test

* fix test gemm multi d

* fix build for weight preshuffle

* fix grouped gemm test

* fix grouped gemm multi d test

* fix grouped gemm preshuffle

* fix grouped gemm example except for quant

* fix gemm preshuffle

* fix splitk 2 stage example

* fix batched gemm example

* fix multid example

* fix multiabd example

* fix batched gemm test

* fixup

* fix examples build

* fix grouped gemm test build

* fix smoke builder

* hacky poc

* fix tile engine

* kill the lambda

* maybe fix test build

* more fixes

* clang-format

* save temp

* clang-format

* mostly fix examples

* clang-format

* remove dead code

* more cleanup

* fix fmha bwd build (default epilogue set/add appears to be broken)

* fix default epilogue tests but not correctness

* clang-format

* fix bquant

* clang-format

* cleanup dead code

* rearrange make windows for readability

* restore changes to IsSupportedArgument

* fix smoke-builder

* clang-format

* fixup rename class

* build fixes

* clang-format

* fix builder

* fixup

* remove set from builder tests

* fix test

* clang-format

* re-refactor the kernels

* clang-format

* fix header license

* remove memory operation from conv bwd test

* clang-format

* clang-format example,include

* clang-format test

* build fixes

* clang-format

* solve compilation error

* fix the CI

* solve compilation error

* clang format

* solve merge conflict

* solve merge conflict

* solve the gfx11 error

* solve test error

* moar build fixes

* remove AtomicAddRequiresKBatchGreaterThanOne test since the property is removed from the kernel scope

---------

Co-authored-by: Thomas Ning <Thomas.Ning@amd.com>

[ROCm/composable_kernel commit: e339101e9c]
This commit is contained in:
Max Podkorytov
2026-01-04 03:28:14 -08:00
committed by GitHub
parent 077d75cea0
commit 6cf89bbca9
68 changed files with 4198 additions and 4298 deletions

View File

@@ -92,67 +92,59 @@ float batched_contraction_impl(const ck_tile::BatchedContractionHostArgs<DsDataT
constexpr auto scheduler = GEMM_PIPELINE_SCHEDULER;
const auto Run = [&]() {
constexpr auto memory_operation =
ck_tile::memory_operation_enum::set; // Always set (no atomic_add)
using UniversalGemmProblem = ck_tile::UniversalGemmPipelineProblem<ADataType,
BDataType,
AccDataType,
GemmShape,
GemmUniversalTraits,
scheduler>;
using UniversalGemmProblem = ck_tile::UniversalGemmPipelineProblem<ADataType,
BDataType,
AccDataType,
GemmShape,
GemmUniversalTraits,
scheduler>;
using GemmPipeline = GEMM_PIPELINE<UniversalGemmProblem>;
using GemmPipeline = GEMM_PIPELINE<UniversalGemmProblem>;
using GemmEpilogue = ck_tile::CShuffleEpilogue<
ck_tile::CShuffleEpilogueProblem<ADataType,
BDataType,
DsDataType,
AccDataType,
EDataType,
DsLayout,
ELayout,
CDEElementWise,
TilePartitioner::MPerBlock,
TilePartitioner::NPerBlock,
M_Warp,
N_Warp,
M_Warp_Tile,
N_Warp_Tile,
K_Warp_Tile,
UniversalGemmProblem::TransposeC>>;
using GemmEpilogue = ck_tile::CShuffleEpilogue<
ck_tile::CShuffleEpilogueProblem<ADataType,
BDataType,
DsDataType,
AccDataType,
EDataType,
DsLayout,
ELayout,
CDEElementWise,
TilePartitioner::MPerBlock,
TilePartitioner::NPerBlock,
M_Warp,
N_Warp,
M_Warp_Tile,
N_Warp_Tile,
K_Warp_Tile,
UniversalGemmProblem::TransposeC,
memory_operation>>;
using Kernel =
ck_tile::BatchedContractionKernel<Problem, TilePartitioner, GemmPipeline, GemmEpilogue>;
auto kargs = Kernel::MakeKernelArgs(args);
using Kernel =
ck_tile::BatchedContractionKernel<Problem, TilePartitioner, GemmPipeline, GemmEpilogue>;
auto kargs = Kernel::MakeKernelArgs(args);
const dim3 grids = Kernel::GridSize(kargs);
const dim3 blocks = Kernel::GetBlockSize();
const dim3 grids = Kernel::GridSize(kargs);
const dim3 blocks = Kernel::GetBlockSize();
if(!Kernel::IsSupportedArguments(kargs))
{
throw std::runtime_error("Wrong! Arguments not supported! Skipping contraction!\n");
}
if(!Kernel::IsSupportedArguments(kargs))
{
throw std::runtime_error("Wrong! Arguments not supported! Skipping contraction!\n");
}
if(s.log_level_ > 0)
{
std::cout << "Launching kernel with args: " << Kernel::GetKernelName() << '\n'
<< "shape: " << GemmShape::GetName() << '\n'
<< "problem: " << GemmPipelineProblem::GetName() << '\n'
<< "pipeline: " << GemmPipeline::GetName() << '\n'
<< "grid: {" << grids.x << ", " << grids.y << ", " << grids.z << "}"
<< ", blocks: {" << blocks.x << ", " << blocks.y << ", " << blocks.z << "}"
<< std::endl;
}
if(s.log_level_ > 0)
{
std::cout << "Launching kernel with args: " << Kernel::GetKernelName() << '\n'
<< "shape: " << GemmShape::GetName() << '\n'
<< "problem: " << GemmPipelineProblem::GetName() << '\n'
<< "pipeline: " << GemmPipeline::GetName() << '\n'
<< "grid: {" << grids.x << ", " << grids.y << ", " << grids.z << "}"
<< ", blocks: {" << blocks.x << ", " << blocks.y << ", " << blocks.z << "}"
<< std::endl;
}
auto kernel = ck_tile::make_kernel<kBlockPerCu>(Kernel{}, grids, blocks, 0, kargs);
auto kernel = ck_tile::make_kernel<kBlockPerCu>(Kernel{}, grids, blocks, 0, kargs);
return ck_tile::launch_kernel(s, kernel);
};
return Run();
return ck_tile::launch_kernel(s, kernel);
}
#define HANDLE_CASE(G, M, N, K) \