Skip to content

GitLab

Commit 20475412 authored by wangshaojie6's avatar wangshaojie6
Browse files

try 4 prefetch

parent b28e7682
Hide whitespace changes
Inline Side-by-side
Showing with 2 additions and 1 deletion
example/01_gemm/gemm_xdl_fp16_splitk.cpp
View file @ 20475412
...@@ -50,7 +50,8 @@ using DeviceGemmInstance = ck::tensor_operation::device::DeviceGemmXdlSplitKCShu ...@@ -50,7 +50,8 @@ using DeviceGemmInstance = ck::tensor_operation::device::DeviceGemmXdlSplitKCShu
//######| | | | Type| Type| Type| Type| DataType| Elementwise| Elementwise| Elementwise| Spacialization| Prefetch| Size| Block| Block| Block| | | XDL| XDL| Per| Per| ThreadCluster| ThreadCluster| SrcAccessOrder| SrcVectorDim| SrcScalar| DstScalar| AddExtraM| ThreadCluster| ThreadCluster| SrcAccessOrder| SrcVectorDim| SrcScalar| DstScalar| AddExtraN| MXdlPerWave| NXdlPerWave| _MBlock_MWaveMPerXdl| ScalarPerVector| //######| | | | Type| Type| Type| Type| DataType| Elementwise| Elementwise| Elementwise| Spacialization| Prefetch| Size| Block| Block| Block| | | XDL| XDL| Per| Per| ThreadCluster| ThreadCluster| SrcAccessOrder| SrcVectorDim| SrcScalar| DstScalar| AddExtraM| ThreadCluster| ThreadCluster| SrcAccessOrder| SrcVectorDim| SrcScalar| DstScalar| AddExtraN| MXdlPerWave| NXdlPerWave| _MBlock_MWaveMPerXdl| ScalarPerVector|
//######| | | | | | | | | Operation| Operation| Operation| | Stage| | | | | | | | | Wave| Wave| Lengths_K0_M_K1| ArrangeOrder| | | PerVector| PerVector_K1| | Lengths_K0_N_K1| ArrangeOrder| | | PerVector| PerVector_K1| | PerShuffle| PerShuffle| _NBlock_NWaveNPerXdl| _NWaveNPerXdl| //######| | | | | | | | | Operation| Operation| Operation| | Stage| | | | | | | | | Wave| Wave| Lengths_K0_M_K1| ArrangeOrder| | | PerVector| PerVector_K1| | Lengths_K0_N_K1| ArrangeOrder| | | PerVector| PerVector_K1| | PerShuffle| PerShuffle| _NBlock_NWaveNPerXdl| _NWaveNPerXdl|
//######| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | //######| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | |
<Row, Row, Row, F16, F16, F16, F32, F16, AElementOp, BElementOp, CElementOp, GemmDefault, 3, 256, 16, 128, 32, 8, 2, 16, 16, 1, 2, S<1, 4, 16, 4>, S<0, 2, 1, 3>, S<0, 2, 1, 3>, 3, 2, 2, 1, S<1, 8, 32, 1>, S<0, 1, 3, 2>, S<0, 1, 3, 2>, 2, 4, 2, 8, 1, 1, S<1, 16, 1, 16>, 2>; //<Row, Row, Row, F16, F16, F16, F32, F16, AElementOp, BElementOp, CElementOp, GemmDefault, 3, 256, 16, 128, 32, 8, 2, 16, 16, 1, 2, S<1, 4, 16, 4>, S<0, 2, 1, 3>, S<0, 2, 1, 3>, 3, 2, 2, 1, S<1, 8, 32, 1>, S<0, 1, 3, 2>, S<0, 1, 3, 2>, 2, 4, 2, 8, 1, 1, S<1, 16, 1, 16>, 2>;
<Row, Row, Row, F16, F16, F16, F32, F16, AElementOp, BElementOp, CElementOp, GemmDefault, 3, 256, 16, 128, 32, 8, 2, 16, 16, 1, 2, S<1, 4, 16, 4>, S<0, 2, 1, 3>, S<0, 2, 1, 3>, 4, 2, 2, 1, S<1, 8, 32, 1>, S<0, 1, 3, 2>, S<0, 1, 3, 2>, 2, 4, 2, 8, 1, 1, S<1, 16, 1, 16>, 2>;
// clang-format on // clang-format on
......
Markdown is supported
0% or .
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment