Skip to content

Commit 2b2a39b

Browse files
vpietila-amdassistant-librarian[bot]
vpietila-amd
authored and
assistant-librarian[bot]
committed
[rocm-libraries] ROCm/rocm-libraries#4275 (commit 2e07a39)
[CK] Add new fwd conv fp16/bf16 instances optimized for unit group size. (#4275) MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit ## Proposed changes Added new FP16/BF16 instances that are optimized for group size = 1. The new instance use the compute optimized block GEMM pipeline. | CK prof command | Baseline (TFLOPs) | New V3 instances (TFLOPs) | |:-----|:------:|------:| | grouped_conv_fwd 1 1 1 0 1 0 1 2 1 32 2376 256 3 3 100 100 1 1 1 1 1 1 1 1 | 858.818 | 962.293 | | grouped_conv_fwd 1 1 1 0 1 0 1 2 1 32 256 256 3 3 100 100 1 1 1 1 1 1 1 1 | 979.987 | 1121.11 | | grouped_conv_fwd 1 1 1 0 1 0 1 2 1 32 2376 256 3 3 50 50 1 1 1 1 1 1 1 1 | 945.951 | 1091.66 |
1 parent 270b144 commit 2b2a39b

1 file changed

Lines changed: 14 additions & 2 deletions

File tree

‎library/include/ck/library/tensor_operation_instance/gpu/grouped_conv_fwd/device_grouped_conv_fwd_xdl_comp_instance.hpp‎

Lines changed: 14 additions & 2 deletions
Original file line numberDiff line numberDiff line change
@@ -63,11 +63,17 @@ template <index_t NDimSpatial,
6363
typename DsDataTypes = Tuple<>,
6464
typename OutElementOp = PassThrough>
6565
using device_grouped_conv_fwd_xdl_bf16_comp_instances_2x = std::tuple<
66-
// clang-format off
66+
// clang-format off
6767
//########################################| NumDim| A| B| Ds| E| AData| BData| AccData| CShuffle| Ds| EData| A| B| CDE| ConvForward| GEMM| NumGemmK| Block| MPer| NPer| KPer| AK1| BK1| MPer| NPer| MXdl| NXdl| ABlockTransfer| ABlockTransfer| ABlockTransfer| ABlockTransfer| ABlockTransfer| ABlockTransfer| ABlockLds| BBlockTransfer| BBlockTransfer| BBlockTransfer| BlockTransfer| BBlockTransfer| BBlockTransfer| BBlockLds| CShuffle| CShuffle| CBlockTransferClusterLengths| CBlockTransfer|
6868
//########################################| Spatial| Layout| Layout| Layout| Layout| Type| Type| Type| DataType| DataType| Type| Elementwise| Elementwise| Elementwise| Specialization| Specialization| 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|
6969
//########################################| | | | | | | | | | | | 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|
7070
//########################################| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | |
71+
#if defined(CK_USE_GFX950)
72+
// Instances optimized for G=1
73+
// The use more shared memory than what is available for non gfx950 architectures.
74+
DeviceGroupedConvFwdMultipleABD_Xdl_CShuffle_V3<NDimSpatial,ALayout,BLayout, DsLayout,ELayout, BF16, BF16, F32, BF16, DsDataTypes, BF16, PassThrough, PassThrough, OutElementOp, ConvSpec, GemmMNKPadding, 256, 256, 256, 64, 8, 8, 32, 32, 4, 4, S<4, 64, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, 1, S<4, 64, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, 1, 1, 1, S<1, 32, 1, 8>, 8, BlockGemmPipelineScheduler::Intrawave, BlockGemmPipelineVersion::v3>,
75+
DeviceGroupedConvFwdMultipleABD_Xdl_CShuffle_V3<NDimSpatial,ALayout,BLayout, DsLayout,ELayout, BF16, BF16, F32, BF16, DsDataTypes, BF16, PassThrough, PassThrough, OutElementOp, ConvSpec, GemmMNKPadding, 256, 512, 128, 32, 8, 8, 32, 32, 8, 2, S<4, 64, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, 1, S<4, 64, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, 1, 1, 1, S<1, 32, 1, 8>, 8, BlockGemmPipelineScheduler::Intrawave, BlockGemmPipelineVersion::v3>,
76+
#endif
7177
DeviceGroupedConvFwdMultipleABD_Xdl_CShuffle_V3<NDimSpatial,ALayout,BLayout, DsLayout,ELayout, BF16, BF16, F32, BF16, DsDataTypes, BF16, PassThrough, PassThrough, OutElementOp, ConvSpec, GemmMNKPadding, 256, 128, 128, 64, 16, 16, 32, 32, 2, 2, S<4, 64, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, 1, S<4, 64, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, 1, 1, 1, S<1, 32, 1, 8>, 8, BlockGemmPipelineScheduler::Interwave, BlockGemmPipelineVersion::v1>
7278
// clang-format on
7379
>;
@@ -131,11 +137,17 @@ template <index_t NDimSpatial,
131137
typename DsDataTypes = Tuple<>,
132138
typename OutElementOp = PassThrough>
133139
using device_grouped_conv_fwd_xdl_f16_comp_instances_2x = std::tuple<
134-
// clang-format off
140+
// clang-format off
135141
//########################################| NumDim| A| B| Ds| E| AData| BData| AccData| CShuffle| Ds| EData| A| B| CDE| ConvForward| GEMM| NumGemmK| Block| MPer| NPer| KPer| AK1| BK1| MPer| NPer| MXdl| NXdl| ABlockTransfer| ABlockTransfer| ABlockTransfer| ABlockTransfer| ABlockTransfer| ABlockTransfer| ABlockLds| BBlockTransfer| BBlockTransfer| BBlockTransfer| BlockTransfer| BBlockTransfer| BBlockTransfer| BBlockLds| CShuffle| CShuffle| CBlockTransferClusterLengths| CBlockTransfer|
136142
//########################################| Spatial| Layout| Layout| Layout| Layout| Type| Type| Type| DataType| DataType| Type| Elementwise| Elementwise| Elementwise| Specialization| Specialization| 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|
137143
//########################################| | | | | | | | | | | | 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|
138144
//########################################| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | |
145+
#if defined(CK_USE_GFX950)
146+
// Instances optimized for G=1
147+
// The use more shared memory than what is available for non gfx950 architectures.
148+
DeviceGroupedConvFwdMultipleABD_Xdl_CShuffle_V3<NDimSpatial,ALayout,BLayout, DsLayout,ELayout, F16, F16, F32, F16, DsDataTypes, F16, PassThrough, PassThrough, OutElementOp, ConvSpec, GemmMNKPadding, 256, 256, 256, 64, 8, 8, 32, 32, 4, 4, S<4, 64, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, 1, S<4, 64, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, 1, 1, 1, S<1, 32, 1, 8>, 8, BlockGemmPipelineScheduler::Intrawave, BlockGemmPipelineVersion::v3>,
149+
DeviceGroupedConvFwdMultipleABD_Xdl_CShuffle_V3<NDimSpatial,ALayout,BLayout, DsLayout,ELayout, F16, F16, F32, F16, DsDataTypes, F16, PassThrough, PassThrough, OutElementOp, ConvSpec, GemmMNKPadding, 256, 512, 128, 32, 8, 8, 32, 32, 8, 2, S<4, 64, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, 1, S<4, 64, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, 1, 1, 1, S<1, 32, 1, 8>, 8, BlockGemmPipelineScheduler::Intrawave, BlockGemmPipelineVersion::v3>,
150+
#endif
139151
DeviceGroupedConvFwdMultipleABD_Xdl_CShuffle_V3<NDimSpatial,ALayout,BLayout, DsLayout,ELayout, F16, F16, F32, F16, DsDataTypes, F16, PassThrough, PassThrough, OutElementOp, ConvSpec, GemmMNKPadding, 256, 128, 128, 64, 16, 16, 32, 32, 2, 2, S<4, 64, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, 1, S<4, 64, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, 1, 1, 1, S<1, 32, 1, 8>, 8, BlockGemmPipelineScheduler::Interwave, BlockGemmPipelineVersion::v1>
140152
// clang-format on
141153
>;

0 commit comments

Comments
 (0)