ROCm/composable_kernel
1
0
Fork 0
mirror of https://github.com/ROCm/composable_kernel.git synced 2026-04-20 14:59:17 +00:00
Code Issues Packages Projects Releases Wiki Activity

[CK_TILE MOE] add NT & preshuffle permute to cktile MOE (#3377)

Browse Source 
* update coherence
---------

Co-authored-by: Zzz9990 <Zzz9990>
This commit is contained in:
Zzz9990
2025-12-10 10:03:28 +08:00
committed by GitHub
parent 934ba1208a
commit 1aa93ef551
8 changed files with 88 additions and 29 deletions
Download Patch File Download Diff File Expand all files Collapse all files

13
include/ck_tile/ops/gemm/pipeline/gemm_pipeline_problem.hpp
View File

@@ -316,10 +316,12 @@ template <typename ADataType_,
typename CDataType_,
typename BlockGemmShape_,
typename Traits_,
GemmPipelineScheduler Scheduler_ = GemmPipelineScheduler::Intrawave,
bool HasHotLoop_ = true,
TailNumber TailNum_ = TailNumber::Full,
typename ComputeDataType_ = ADataType_>
GemmPipelineScheduler Scheduler_ = GemmPipelineScheduler::Intrawave,
bool HasHotLoop_ = true,
TailNumber TailNum_ = TailNumber::Full,
amd_buffer_coherence_enum BMemNTType_ = amd_buffer_coherence_enum::coherence_default,
bool BPreShufflePermute_ = false,
typename ComputeDataType_ = ADataType_>
struct FlatmmPipelineProblem
{
using Traits = remove_cvref_t<Traits_>;
@@ -353,6 +355,9 @@ struct FlatmmPipelineProblem
static constexpr auto HasHotLoop = HasHotLoop_;
static constexpr auto TailNum = TailNum_;
static constexpr auto BMemNTType = BMemNTType_;
static constexpr bool BPreShufflePermute = BPreShufflePermute_;
[[nodiscard]] CK_TILE_HOST static const std::string GetName()
{
// clang-format off