Skip to content

GitLab

Commit 9292361d authored by aska-0096's avatar aska-0096
Browse files

tunning kernel

parent df7a58bc
Hide whitespace changes
Inline Side-by-side
Showing with 125 additions and 124 deletions
example/30_grouped_conv_fwd_multiple_d/run_grouped_conv_fwd_bias_relu_add_wmma_example.inc
View file @ 9292361d
......@@ -47,7 +47,7 @@ using DeviceConvFwdInstances = std::tuple<
2, 1, // MRepeat x NRepeat
S<1, 256, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 16, 16, true, // ABlockTransferThreadClusterLengths_AK0_M_AK1, ArrangeOrder, SrcAccessOrder, VectorDim, SrcScalarPerVector, DstScalarPerVector_AK1, LdsExtraM
S<1, 16, 8>, S<1, 0, 2>, S<1, 0, 2>, 2, 2, 2, true,
1, 1, S<1, 256, 1, 1>, 16>,
2, 1, S<1, 256, 1, 1>, 16>,
ck::tensor_operation::device::DeviceGroupedConvFwdMultipleD_Wmma_CShuffle<
NDimSpatial,
InputLayout<NDimSpatial>, WeightLayout<NDimSpatial>, ck::Tuple<BiasLayout<NDimSpatial>, ResidualLayout<NDimSpatial>>, OutputLayout<NDimSpatial>,
......@@ -58,7 +58,7 @@ using DeviceConvFwdInstances = std::tuple<
2, 1, // MRepeat x NRepeat
S<1, 128, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 16, 16, true, // ABlockTransferThreadClusterLengths_AK0_M_AK1, ArrangeOrder, SrcAccessOrder, VectorDim, SrcScalarPerVector, DstScalarPerVector_AK1, LdsExtraM
S<1, 16, 8>, S<1, 0, 2>, S<1, 0, 2>, 2, 2, 2, true,
1, 1, S<1, 128, 1, 1>, 16>,
2, 1, S<1, 128, 1, 1>, 16>,
ck::tensor_operation::device::DeviceGroupedConvFwdMultipleD_Wmma_CShuffle<
NDimSpatial,
InputLayout<NDimSpatial>, WeightLayout<NDimSpatial>, ck::Tuple<BiasLayout<NDimSpatial>, ResidualLayout<NDimSpatial>>, OutputLayout<NDimSpatial>,
......@@ -69,7 +69,7 @@ using DeviceConvFwdInstances = std::tuple<
2, 1, // MRepeat x NRepeat
S<1, 64, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 16, 16, true, // ABlockTransferThreadClusterLengths_AK0_M_AK1, ArrangeOrder, SrcAccessOrder, VectorDim, SrcScalarPerVector, DstScalarPerVector_AK1, LdsExtraM
S<1, 16, 4>, S<1, 0, 2>, S<1, 0, 2>, 2, 4, 4, true,
1, 1, S<1, 64, 1, 1>, 16>,
2, 1, S<1, 64, 1, 1>, 16>,
// K0 = 3
ck::tensor_operation::device::DeviceGroupedConvFwdMultipleD_Wmma_CShuffle<
NDimSpatial,
......@@ -81,7 +81,7 @@ using DeviceConvFwdInstances = std::tuple<
2, 1, // MRepeat x NRepeat
S<1, 256, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 16, 16, true, // ABlockTransferThreadClusterLengths_AK0_M_AK1, ArrangeOrder, SrcAccessOrder, VectorDim, SrcScalarPerVector, DstScalarPerVector_AK1, LdsExtraM
S<1, 16, 8>, S<1, 0, 2>, S<1, 0, 2>, 2, 2, 2, true,
1, 1, S<1, 256, 1, 1>, 16>,
2, 1, S<1, 256, 1, 1>, 16>,
ck::tensor_operation::device::DeviceGroupedConvFwdMultipleD_Wmma_CShuffle<
NDimSpatial,
InputLayout<NDimSpatial>, WeightLayout<NDimSpatial>, ck::Tuple<BiasLayout<NDimSpatial>, ResidualLayout<NDimSpatial>>, OutputLayout<NDimSpatial>,
......@@ -92,7 +92,7 @@ using DeviceConvFwdInstances = std::tuple<
2, 1, // MRepeat x NRepeat
S<1, 192, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 16, 16, true, // ABlockTransferThreadClusterLengths_AK0_M_AK1, ArrangeOrder, SrcAccessOrder, VectorDim, SrcScalarPerVector, DstScalarPerVector_AK1, LdsExtraM
S<3, 16, 4>, S<1, 0, 2>, S<1, 0, 2>, 2, 4, 4, true,
1, 1, S<1, 64, 1, 3>, 16>,
2, 1, S<1, 192, 1, 1>, 16>,
ck::tensor_operation::device::DeviceGroupedConvFwdMultipleD_Wmma_CShuffle<
NDimSpatial,
InputLayout<NDimSpatial>, WeightLayout<NDimSpatial>, ck::Tuple<BiasLayout<NDimSpatial>, ResidualLayout<NDimSpatial>>, OutputLayout<NDimSpatial>,
......@@ -103,7 +103,7 @@ using DeviceConvFwdInstances = std::tuple<
2, 1, // MRepeat x NRepeat
S<1, 128, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 16, 16, true, // ABlockTransferThreadClusterLengths_AK0_M_AK1, ArrangeOrder, SrcAccessOrder, VectorDim, SrcScalarPerVector, DstScalarPerVector_AK1, LdsExtraM
S<1, 16, 8>, S<1, 0, 2>, S<1, 0, 2>, 2, 2, 2, true,
1, 1, S<1, 128, 1, 1>, 16>,
2, 1, S<1, 128, 1, 1>, 16>,
ck::tensor_operation::device::DeviceGroupedConvFwdMultipleD_Wmma_CShuffle<
NDimSpatial,
InputLayout<NDimSpatial>, WeightLayout<NDimSpatial>, ck::Tuple<BiasLayout<NDimSpatial>, ResidualLayout<NDimSpatial>>, OutputLayout<NDimSpatial>,
......@@ -114,7 +114,7 @@ using DeviceConvFwdInstances = std::tuple<
2, 1, // MRepeat x NRepeat
S<1, 64, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 16, 16, true, // ABlockTransferThreadClusterLengths_AK0_M_AK1, ArrangeOrder, SrcAccessOrder, VectorDim, SrcScalarPerVector, DstScalarPerVector_AK1, LdsExtraM
S<1, 16, 4>, S<1, 0, 2>, S<1, 0, 2>, 2, 4, 4, true,
1, 1, S<1, 64, 1, 1>, 16>,
2, 1, S<1, 64, 1, 1>, 16>,
// K0 = 4
ck::tensor_operation::device::DeviceGroupedConvFwdMultipleD_Wmma_CShuffle<
NDimSpatial,
......@@ -126,7 +126,7 @@ using DeviceConvFwdInstances = std::tuple<
2, 1, // MRepeat x NRepeat
S<1, 256, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 16, 16, true, // ABlockTransferThreadClusterLengths_AK0_M_AK1, ArrangeOrder, SrcAccessOrder, VectorDim, SrcScalarPerVector, DstScalarPerVector_AK1, LdsExtraM
S<1, 16, 8>, S<1, 0, 2>, S<1, 0, 2>, 2, 2, 2, true,
1, 1, S<1, 256, 1, 1>, 16>,
2, 1, S<1, 256, 1, 1>, 16>,
ck::tensor_operation::device::DeviceGroupedConvFwdMultipleD_Wmma_CShuffle<
NDimSpatial,
InputLayout<NDimSpatial>, WeightLayout<NDimSpatial>, ck::Tuple<BiasLayout<NDimSpatial>, ResidualLayout<NDimSpatial>>, OutputLayout<NDimSpatial>,
......@@ -137,7 +137,7 @@ using DeviceConvFwdInstances = std::tuple<
2, 1, // MRepeat x NRepeat
S<1, 128, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 16, 16, true, // ABlockTransferThreadClusterLengths_AK0_M_AK1, ArrangeOrder, SrcAccessOrder, VectorDim, SrcScalarPerVector, DstScalarPerVector_AK1, LdsExtraM
S<1, 16, 8>, S<1, 0, 2>, S<1, 0, 2>, 2, 2, 2, true,
1, 1, S<1, 128, 1, 1>, 16>,
2, 1, S<1, 128, 1, 1>, 16>,
ck::tensor_operation::device::DeviceGroupedConvFwdMultipleD_Wmma_CShuffle<
NDimSpatial,
InputLayout<NDimSpatial>, WeightLayout<NDimSpatial>, ck::Tuple<BiasLayout<NDimSpatial>, ResidualLayout<NDimSpatial>>, OutputLayout<NDimSpatial>,
......@@ -148,7 +148,7 @@ using DeviceConvFwdInstances = std::tuple<
2, 1, // MRepeat x NRepeat
S<1, 64, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 16, 16, true, // ABlockTransferThreadClusterLengths_AK0_M_AK1, ArrangeOrder, SrcAccessOrder, VectorDim, SrcScalarPerVector, DstScalarPerVector_AK1, LdsExtraM
S<1, 16, 4>, S<1, 0, 2>, S<1, 0, 2>, 2, 4, 4, true,
1, 1, S<1, 64, 1, 1>, 16>,
2, 1, S<1, 64, 1, 1>, 16>,
// K0 = 5
ck::tensor_operation::device::DeviceGroupedConvFwdMultipleD_Wmma_CShuffle<
NDimSpatial,
......@@ -160,7 +160,7 @@ using DeviceConvFwdInstances = std::tuple<
2, 1, // MRepeat x NRepeat
S<1, 256, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 16, 16, true, // ABlockTransferThreadClusterLengths_AK0_M_AK1, ArrangeOrder, SrcAccessOrder, VectorDim, SrcScalarPerVector, DstScalarPerVector_AK1, LdsExtraM
S<1, 16, 8>, S<1, 0, 2>, S<1, 0, 2>, 2, 2, 2, true,
1, 1, S<1, 256, 1, 1>, 16>,
2, 1, S<1, 256, 1, 1>, 16>,
ck::tensor_operation::device::DeviceGroupedConvFwdMultipleD_Wmma_CShuffle<
NDimSpatial,
InputLayout<NDimSpatial>, WeightLayout<NDimSpatial>, ck::Tuple<BiasLayout<NDimSpatial>, ResidualLayout<NDimSpatial>>, OutputLayout<NDimSpatial>,
......@@ -171,7 +171,7 @@ using DeviceConvFwdInstances = std::tuple<
2, 1, // MRepeat x NRepeat
S<1, 128, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 16, 16, true, // ABlockTransferThreadClusterLengths_AK0_M_AK1, ArrangeOrder, SrcAccessOrder, VectorDim, SrcScalarPerVector, DstScalarPerVector_AK1, LdsExtraM
S<1, 16, 8>, S<1, 0, 2>, S<1, 0, 2>, 2, 2, 2, true,
1, 1, S<1, 128, 1, 1>, 16>,
2, 1, S<1, 128, 1, 1>, 16>,
ck::tensor_operation::device::DeviceGroupedConvFwdMultipleD_Wmma_CShuffle<
NDimSpatial,
InputLayout<NDimSpatial>, WeightLayout<NDimSpatial>, ck::Tuple<BiasLayout<NDimSpatial>, ResidualLayout<NDimSpatial>>, OutputLayout<NDimSpatial>,
......@@ -182,7 +182,7 @@ using DeviceConvFwdInstances = std::tuple<
2, 1, // MRepeat x NRepeat
S<1, 64, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 16, 16, true, // ABlockTransferThreadClusterLengths_AK0_M_AK1, ArrangeOrder, SrcAccessOrder, VectorDim, SrcScalarPerVector, DstScalarPerVector_AK1, LdsExtraM
S<1, 16, 4>, S<1, 0, 2>, S<1, 0, 2>, 2, 4, 4, true,
1, 1, S<1, 64, 1, 1>, 16>,
2, 1, S<1, 64, 1, 1>, 16>,
// K0 = 6
ck::tensor_operation::device::DeviceGroupedConvFwdMultipleD_Wmma_CShuffle<
NDimSpatial,
......@@ -194,7 +194,7 @@ using DeviceConvFwdInstances = std::tuple<
2, 1, // MRepeat x NRepeat
S<1, 256, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 16, 16, true, // ABlockTransferThreadClusterLengths_AK0_M_AK1, ArrangeOrder, SrcAccessOrder, VectorDim, SrcScalarPerVector, DstScalarPerVector_AK1, LdsExtraM
S<1, 16, 8>, S<1, 0, 2>, S<1, 0, 2>, 2, 2, 2, true,
1, 1, S<1, 256, 1, 1>, 16>,
2, 1, S<1, 256, 1, 1>, 16>,
ck::tensor_operation::device::DeviceGroupedConvFwdMultipleD_Wmma_CShuffle<
NDimSpatial,
InputLayout<NDimSpatial>, WeightLayout<NDimSpatial>, ck::Tuple<BiasLayout<NDimSpatial>, ResidualLayout<NDimSpatial>>, OutputLayout<NDimSpatial>,
......@@ -205,7 +205,7 @@ using DeviceConvFwdInstances = std::tuple<
2, 1, // MRepeat x NRepeat
S<1, 128, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 16, 16, true, // ABlockTransferThreadClusterLengths_AK0_M_AK1, ArrangeOrder, SrcAccessOrder, VectorDim, SrcScalarPerVector, DstScalarPerVector_AK1, LdsExtraM
S<1, 16, 8>, S<1, 0, 2>, S<1, 0, 2>, 2, 2, 2, true,
1, 1, S<1, 128, 1, 1>, 16>,
2, 1, S<1, 128, 1, 1>, 16>,
ck::tensor_operation::device::DeviceGroupedConvFwdMultipleD_Wmma_CShuffle<
NDimSpatial,
InputLayout<NDimSpatial>, WeightLayout<NDimSpatial>, ck::Tuple<BiasLayout<NDimSpatial>, ResidualLayout<NDimSpatial>>, OutputLayout<NDimSpatial>,
......@@ -216,7 +216,7 @@ using DeviceConvFwdInstances = std::tuple<
2, 1, // MRepeat x NRepeat
S<1, 64, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 16, 16, true, // ABlockTransferThreadClusterLengths_AK0_M_AK1, ArrangeOrder, SrcAccessOrder, VectorDim, SrcScalarPerVector, DstScalarPerVector_AK1, LdsExtraM
S<1, 16, 4>, S<1, 0, 2>, S<1, 0, 2>, 2, 4, 4, true,
1, 1, S<1, 64, 1, 1>, 16>,
2, 1, S<1, 64, 1, 1>, 16>,
// K0 = 7
ck::tensor_operation::device::DeviceGroupedConvFwdMultipleD_Wmma_CShuffle<
NDimSpatial,
......@@ -228,7 +228,7 @@ using DeviceConvFwdInstances = std::tuple<
2, 1, // MRepeat x NRepeat
S<1, 256, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 16, 16, true, // ABlockTransferThreadClusterLengths_AK0_M_AK1, ArrangeOrder, SrcAccessOrder, VectorDim, SrcScalarPerVector, DstScalarPerVector_AK1, LdsExtraM
S<1, 16, 8>, S<1, 0, 2>, S<1, 0, 2>, 2, 2, 2, true,
1, 1, S<1, 256, 1, 1>, 16>,
2, 1, S<1, 256, 1, 1>, 16>,
ck::tensor_operation::device::DeviceGroupedConvFwdMultipleD_Wmma_CShuffle<
NDimSpatial,
InputLayout<NDimSpatial>, WeightLayout<NDimSpatial>, ck::Tuple<BiasLayout<NDimSpatial>, ResidualLayout<NDimSpatial>>, OutputLayout<NDimSpatial>,
......@@ -239,7 +239,7 @@ using DeviceConvFwdInstances = std::tuple<
2, 1, // MRepeat x NRepeat
S<1, 128, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 16, 16, true, // ABlockTransferThreadClusterLengths_AK0_M_AK1, ArrangeOrder, SrcAccessOrder, VectorDim, SrcScalarPerVector, DstScalarPerVector_AK1, LdsExtraM
S<1, 16, 8>, S<1, 0, 2>, S<1, 0, 2>, 2, 2, 2, true,
1, 1, S<1, 128, 1, 1>, 16>,
2, 1, S<1, 128, 1, 1>, 16>,
ck::tensor_operation::device::DeviceGroupedConvFwdMultipleD_Wmma_CShuffle<
NDimSpatial,
InputLayout<NDimSpatial>, WeightLayout<NDimSpatial>, ck::Tuple<BiasLayout<NDimSpatial>, ResidualLayout<NDimSpatial>>, OutputLayout<NDimSpatial>,
......@@ -250,7 +250,7 @@ using DeviceConvFwdInstances = std::tuple<
2, 1, // MRepeat x NRepeat
S<1, 64, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 16, 16, true, // ABlockTransferThreadClusterLengths_AK0_M_AK1, ArrangeOrder, SrcAccessOrder, VectorDim, SrcScalarPerVector, DstScalarPerVector_AK1, LdsExtraM
S<1, 16, 4>, S<1, 0, 2>, S<1, 0, 2>, 2, 4, 4, true,
1, 1, S<1, 64, 1, 1>, 16>,
2, 1, S<1, 64, 1, 1>, 16>,
// K0 = 8
ck::tensor_operation::device::DeviceGroupedConvFwdMultipleD_Wmma_CShuffle<
NDimSpatial,
......@@ -262,7 +262,7 @@ using DeviceConvFwdInstances = std::tuple<
2, 1, // MRepeat x NRepeat
S<1, 256, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 16, 16, true, // ABlockTransferThreadClusterLengths_AK0_M_AK1, ArrangeOrder, SrcAccessOrder, VectorDim, SrcScalarPerVector, DstScalarPerVector_AK1, LdsExtraM
S<1, 16, 8>, S<1, 0, 2>, S<1, 0, 2>, 2, 2, 2, true,
1, 1, S<1, 256, 1, 1>, 16>,
2, 1, S<1, 256, 1, 1>, 16>,
ck::tensor_operation::device::DeviceGroupedConvFwdMultipleD_Wmma_CShuffle<
NDimSpatial,
InputLayout<NDimSpatial>, WeightLayout<NDimSpatial>, ck::Tuple<BiasLayout<NDimSpatial>, ResidualLayout<NDimSpatial>>, OutputLayout<NDimSpatial>,
......@@ -273,7 +273,7 @@ using DeviceConvFwdInstances = std::tuple<
2, 1, // MRepeat x NRepeat
S<1, 128, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 16, 16, true, // ABlockTransferThreadClusterLengths_AK0_M_AK1, ArrangeOrder, SrcAccessOrder, VectorDim, SrcScalarPerVector, DstScalarPerVector_AK1, LdsExtraM
S<1, 16, 8>, S<1, 0, 2>, S<1, 0, 2>, 2, 2, 2, true,
1, 1, S<1, 128, 1, 1>, 16>,
2, 1, S<1, 128, 1, 1>, 16>,
ck::tensor_operation::device::DeviceGroupedConvFwdMultipleD_Wmma_CShuffle<
NDimSpatial,
InputLayout<NDimSpatial>, WeightLayout<NDimSpatial>, ck::Tuple<BiasLayout<NDimSpatial>, ResidualLayout<NDimSpatial>>, OutputLayout<NDimSpatial>,
......@@ -284,7 +284,7 @@ using DeviceConvFwdInstances = std::tuple<
2, 1, // MRepeat x NRepeat
S<1, 64, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 16, 16, true, // ABlockTransferThreadClusterLengths_AK0_M_AK1, ArrangeOrder, SrcAccessOrder, VectorDim, SrcScalarPerVector, DstScalarPerVector_AK1, LdsExtraM
S<1, 16, 4>, S<1, 0, 2>, S<1, 0, 2>, 2, 4, 4, true,
1, 1, S<1, 64, 1, 1>, 16>,
2, 1, S<1, 64, 1, 1>, 16>,
// GEMM_N=32, Wave N = 1
// K0 = 2
ck::tensor_operation::device::DeviceGroupedConvFwdMultipleD_Wmma_CShuffle<
......@@ -297,7 +297,7 @@ using DeviceConvFwdInstances = std::tuple<
4, 1, // MRepeat x NRepeat
S<1, 256, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 16, 16, true, // ABlockTransferThreadClusterLengths_AK0_M_AK1, ArrangeOrder, SrcAccessOrder, VectorDim, SrcScalarPerVector, DstScalarPerVector_AK1, LdsExtraM
S<1, 32, 8>, S<1, 0, 2>, S<1, 0, 2>, 2, 4, 4, true,
1, 1, S<1, 128, 1, 2>, 16>,
4, 1, S<1, 128, 1, 2>, 16>,
ck::tensor_operation::device::DeviceGroupedConvFwdMultipleD_Wmma_CShuffle<
NDimSpatial,
InputLayout<NDimSpatial>, WeightLayout<NDimSpatial>, ck::Tuple<BiasLayout<NDimSpatial>, ResidualLayout<NDimSpatial>>, OutputLayout<NDimSpatial>,
......@@ -308,7 +308,7 @@ using DeviceConvFwdInstances = std::tuple<
4, 1, // MRepeat x NRepeat
S<1, 128, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 16, 16, true, // ABlockTransferThreadClusterLengths_AK0_M_AK1, ArrangeOrder, SrcAccessOrder, VectorDim, SrcScalarPerVector, DstScalarPerVector_AK1, LdsExtraM
S<1, 32, 4>, S<1, 0, 2>, S<1, 0, 2>, 2, 4, 4, true,
1, 1, S<1, 64, 1, 2>, 16>,
4, 1, S<1, 64, 1, 2>, 16>,
ck::tensor_operation::device::DeviceGroupedConvFwdMultipleD_Wmma_CShuffle<
NDimSpatial,
InputLayout<NDimSpatial>, WeightLayout<NDimSpatial>, ck::Tuple<BiasLayout<NDimSpatial>, ResidualLayout<NDimSpatial>>, OutputLayout<NDimSpatial>,
......@@ -319,7 +319,7 @@ using DeviceConvFwdInstances = std::tuple<
4, 1, // MRepeat x NRepeat
S<1, 64, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 16, 16, true, // ABlockTransferThreadClusterLengths_AK0_M_AK1, ArrangeOrder, SrcAccessOrder, VectorDim, SrcScalarPerVector, DstScalarPerVector_AK1, LdsExtraM
S<1, 32, 2>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, true,
1, 1, S<1, 32, 1, 2>, 16>,
4, 1, S<1, 32, 1, 2>, 16>,
// K0 = 3
ck::tensor_operation::device::DeviceGroupedConvFwdMultipleD_Wmma_CShuffle<
NDimSpatial,
......@@ -331,7 +331,7 @@ using DeviceConvFwdInstances = std::tuple<
4, 1, // MRepeat x NRepeat
S<1, 256, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 16, 16, true, // ABlockTransferThreadClusterLengths_AK0_M_AK1, ArrangeOrder, SrcAccessOrder, VectorDim, SrcScalarPerVector, DstScalarPerVector_AK1, LdsExtraM
S<1, 32, 8>, S<1, 0, 2>, S<1, 0, 2>, 2, 4, 4, true,
1, 1, S<1, 128, 1, 2>, 16>,
4, 1, S<1, 128, 1, 2>, 16>,
ck::tensor_operation::device::DeviceGroupedConvFwdMultipleD_Wmma_CShuffle<
NDimSpatial,
InputLayout<NDimSpatial>, WeightLayout<NDimSpatial>, ck::Tuple<BiasLayout<NDimSpatial>, ResidualLayout<NDimSpatial>>, OutputLayout<NDimSpatial>,
......@@ -342,7 +342,7 @@ using DeviceConvFwdInstances = std::tuple<
4, 1, // MRepeat x NRepeat
S<1, 192, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 16, 16, true, // ABlockTransferThreadClusterLengths_AK0_M_AK1, ArrangeOrder, SrcAccessOrder, VectorDim, SrcScalarPerVector, DstScalarPerVector_AK1, LdsExtraM
S<3, 32, 2>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, true,
1, 1, S<1, 96, 1, 2>, 16>,
4, 1, S<1, 96, 1, 2>, 16>,
ck::tensor_operation::device::DeviceGroupedConvFwdMultipleD_Wmma_CShuffle<
NDimSpatial,
InputLayout<NDimSpatial>, WeightLayout<NDimSpatial>, ck::Tuple<BiasLayout<NDimSpatial>, ResidualLayout<NDimSpatial>>, OutputLayout<NDimSpatial>,
......@@ -353,7 +353,7 @@ using DeviceConvFwdInstances = std::tuple<
4, 1, // MRepeat x NRepeat
S<1, 128, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 16, 16, true, // ABlockTransferThreadClusterLengths_AK0_M_AK1, ArrangeOrder, SrcAccessOrder, VectorDim, SrcScalarPerVector, DstScalarPerVector_AK1, LdsExtraM
S<1, 32, 4>, S<1, 0, 2>, S<1, 0, 2>, 2, 4, 4, true,
1, 1, S<1, 64, 1, 2>, 16>,
4, 1, S<1, 64, 1, 2>, 16>,
ck::tensor_operation::device::DeviceGroupedConvFwdMultipleD_Wmma_CShuffle<
NDimSpatial,
InputLayout<NDimSpatial>, WeightLayout<NDimSpatial>, ck::Tuple<BiasLayout<NDimSpatial>, ResidualLayout<NDimSpatial>>, OutputLayout<NDimSpatial>,
......@@ -364,7 +364,7 @@ using DeviceConvFwdInstances = std::tuple<
4, 1, // MRepeat x NRepeat
S<1, 64, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 16, 16, true, // ABlockTransferThreadClusterLengths_AK0_M_AK1, ArrangeOrder, SrcAccessOrder, VectorDim, SrcScalarPerVector, DstScalarPerVector_AK1, LdsExtraM
S<1, 32, 2>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, true,
1, 1, S<1, 32, 1, 2>, 16>,
4, 1, S<1, 32, 1, 2>, 16>,
// K0 = 4
ck::tensor_operation::device::DeviceGroupedConvFwdMultipleD_Wmma_CShuffle<
NDimSpatial,
......@@ -376,7 +376,7 @@ using DeviceConvFwdInstances = std::tuple<
4, 1, // MRepeat x NRepeat
S<1, 256, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 16, 16, true, // ABlockTransferThreadClusterLengths_AK0_M_AK1, ArrangeOrder, SrcAccessOrder, VectorDim, SrcScalarPerVector, DstScalarPerVector_AK1, LdsExtraM
S<1, 16, 8>, S<1, 0, 2>, S<1, 0, 2>, 2, 2, 2, true,
1, 1, S<1, 128, 1, 2>, 16>,
4, 1, S<1, 128, 1, 2>, 16>,
ck::tensor_operation::device::DeviceGroupedConvFwdMultipleD_Wmma_CShuffle<
NDimSpatial,
InputLayout<NDimSpatial>, WeightLayout<NDimSpatial>, ck::Tuple<BiasLayout<NDimSpatial>, ResidualLayout<NDimSpatial>>, OutputLayout<NDimSpatial>,
......@@ -387,7 +387,7 @@ using DeviceConvFwdInstances = std::tuple<
4, 1, // MRepeat x NRepeat
S<1, 128, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 16, 16, true, // ABlockTransferThreadClusterLengths_AK0_M_AK1, ArrangeOrder, SrcAccessOrder, VectorDim, SrcScalarPerVector, DstScalarPerVector_AK1, LdsExtraM
S<1, 32, 4>, S<1, 0, 2>, S<1, 0, 2>, 2, 4, 4, true,
1, 1, S<1, 64, 1, 2>, 16>,
4, 1, S<1, 64, 1, 2>, 16>,
ck::tensor_operation::device::DeviceGroupedConvFwdMultipleD_Wmma_CShuffle<
NDimSpatial,
InputLayout<NDimSpatial>, WeightLayout<NDimSpatial>, ck::Tuple<BiasLayout<NDimSpatial>, ResidualLayout<NDimSpatial>>, OutputLayout<NDimSpatial>,
......@@ -398,7 +398,7 @@ using DeviceConvFwdInstances = std::tuple<
4, 1, // MRepeat x NRepeat
S<1, 64, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 16, 16, true, // ABlockTransferThreadClusterLengths_AK0_M_AK1, ArrangeOrder, SrcAccessOrder, VectorDim, SrcScalarPerVector, DstScalarPerVector_AK1, LdsExtraM
S<1, 32, 2>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, true,
1, 1, S<1, 32, 1, 2>, 16>,
4, 1, S<1, 32, 1, 2>, 16>,
// K0 = 5
ck::tensor_operation::device::DeviceGroupedConvFwdMultipleD_Wmma_CShuffle<
NDimSpatial,
......@@ -410,7 +410,7 @@ using DeviceConvFwdInstances = std::tuple<
4, 1, // MRepeat x NRepeat
S<1, 256, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 16, 16, true, // ABlockTransferThreadClusterLengths_AK0_M_AK1, ArrangeOrder, SrcAccessOrder, VectorDim, SrcScalarPerVector, DstScalarPerVector_AK1, LdsExtraM
S<1, 32, 8>, S<1, 0, 2>, S<1, 0, 2>, 2, 2, 2, true,
1, 1, S<1, 128, 1, 2>, 16>,
4, 1, S<1, 128, 1, 2>, 16>,
ck::tensor_operation::device::DeviceGroupedConvFwdMultipleD_Wmma_CShuffle<
NDimSpatial,
InputLayout<NDimSpatial>, WeightLayout<NDimSpatial>, ck::Tuple<BiasLayout<NDimSpatial>, ResidualLayout<NDimSpatial>>, OutputLayout<NDimSpatial>,
......@@ -421,7 +421,7 @@ using DeviceConvFwdInstances = std::tuple<
4, 1, // MRepeat x NRepeat
S<1, 128, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 16, 16, true, // ABlockTransferThreadClusterLengths_AK0_M_AK1, ArrangeOrder, SrcAccessOrder, VectorDim, SrcScalarPerVector, DstScalarPerVector_AK1, LdsExtraM
S<1, 32, 4>, S<1, 0, 2>, S<1, 0, 2>, 2, 4, 4, true,
1, 1, S<1, 64, 1, 2>, 16>,
4, 1, S<1, 64, 1, 2>, 16>,
ck::tensor_operation::device::DeviceGroupedConvFwdMultipleD_Wmma_CShuffle<
NDimSpatial,
InputLayout<NDimSpatial>, WeightLayout<NDimSpatial>, ck::Tuple<BiasLayout<NDimSpatial>, ResidualLayout<NDimSpatial>>, OutputLayout<NDimSpatial>,
......@@ -432,7 +432,7 @@ using DeviceConvFwdInstances = std::tuple<
4, 1, // MRepeat x NRepeat
S<1, 64, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 16, 16, true, // ABlockTransferThreadClusterLengths_AK0_M_AK1, ArrangeOrder, SrcAccessOrder, VectorDim, SrcScalarPerVector, DstScalarPerVector_AK1, LdsExtraM
S<1, 32, 2>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, true,
1, 1, S<1, 32, 1, 2>, 16>,
4, 1, S<1, 32, 1, 2>, 16>,
// K0 = 6
ck::tensor_operation::device::DeviceGroupedConvFwdMultipleD_Wmma_CShuffle<
NDimSpatial,
......@@ -444,7 +444,7 @@ using DeviceConvFwdInstances = std::tuple<
4, 1, // MRepeat x NRepeat
S<1, 256, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 16, 16, true, // ABlockTransferThreadClusterLengths_AK0_M_AK1, ArrangeOrder, SrcAccessOrder, VectorDim, SrcScalarPerVector, DstScalarPerVector_AK1, LdsExtraM
S<1, 32, 8>, S<1, 0, 2>, S<1, 0, 2>, 2, 2, 2, true,
1, 1, S<1, 128, 1, 2>, 16>,
4, 1, S<1, 128, 1, 2>, 16>,
ck::tensor_operation::device::DeviceGroupedConvFwdMultipleD_Wmma_CShuffle<
NDimSpatial,
InputLayout<NDimSpatial>, WeightLayout<NDimSpatial>, ck::Tuple<BiasLayout<NDimSpatial>, ResidualLayout<NDimSpatial>>, OutputLayout<NDimSpatial>,
......@@ -455,7 +455,7 @@ using DeviceConvFwdInstances = std::tuple<
4, 1, // MRepeat x NRepeat
S<1, 128, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 16, 16, true, // ABlockTransferThreadClusterLengths_AK0_M_AK1, ArrangeOrder, SrcAccessOrder, VectorDim, SrcScalarPerVector, DstScalarPerVector_AK1, LdsExtraM
S<1, 32, 4>, S<1, 0, 2>, S<1, 0, 2>, 2, 4, 4, true,
1, 1, S<1, 64, 1, 2>, 16>,
4, 1, S<1, 64, 1, 2>, 16>,
ck::tensor_operation::device::DeviceGroupedConvFwdMultipleD_Wmma_CShuffle<
NDimSpatial,
InputLayout<NDimSpatial>, WeightLayout<NDimSpatial>, ck::Tuple<BiasLayout<NDimSpatial>, ResidualLayout<NDimSpatial>>, OutputLayout<NDimSpatial>,
......@@ -466,7 +466,7 @@ using DeviceConvFwdInstances = std::tuple<
4, 1, // MRepeat x NRepeat
S<1, 64, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 16, 16, true, // ABlockTransferThreadClusterLengths_AK0_M_AK1, ArrangeOrder, SrcAccessOrder, VectorDim, SrcScalarPerVector, DstScalarPerVector_AK1, LdsExtraM
S<1, 32, 2>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, true,
1, 1, S<1, 32, 1, 2>, 16>,
4, 1, S<1, 32, 1, 2>, 16>,
// K0 = 7
ck::tensor_operation::device::DeviceGroupedConvFwdMultipleD_Wmma_CShuffle<
NDimSpatial,
......@@ -478,7 +478,7 @@ using DeviceConvFwdInstances = std::tuple<
4, 1, // MRepeat x NRepeat
S<1, 256, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 16, 16, true, // ABlockTransferThreadClusterLengths_AK0_M_AK1, ArrangeOrder, SrcAccessOrder, VectorDim, SrcScalarPerVector, DstScalarPerVector_AK1, LdsExtraM
S<1, 32, 8>, S<1, 0, 2>, S<1, 0, 2>, 2, 2, 2, true,
1, 1, S<1, 128, 1, 2>, 16>,
4, 1, S<1, 128, 1, 2>, 16>,
ck::tensor_operation::device::DeviceGroupedConvFwdMultipleD_Wmma_CShuffle<
NDimSpatial,
InputLayout<NDimSpatial>, WeightLayout<NDimSpatial>, ck::Tuple<BiasLayout<NDimSpatial>, ResidualLayout<NDimSpatial>>, OutputLayout<NDimSpatial>,
......@@ -489,7 +489,7 @@ using DeviceConvFwdInstances = std::tuple<
4, 1, // MRepeat x NRepeat
S<1, 128, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 16, 16, true, // ABlockTransferThreadClusterLengths_AK0_M_AK1, ArrangeOrder, SrcAccessOrder, VectorDim, SrcScalarPerVector, DstScalarPerVector_AK1, LdsExtraM
S<1, 32, 4>, S<1, 0, 2>, S<1, 0, 2>, 2, 4, 4, true,
1, 1, S<1, 64, 1, 2>, 16>,
4, 1, S<1, 64, 1, 2>, 16>,
ck::tensor_operation::device::DeviceGroupedConvFwdMultipleD_Wmma_CShuffle<
NDimSpatial,
InputLayout<NDimSpatial>, WeightLayout<NDimSpatial>, ck::Tuple<BiasLayout<NDimSpatial>, ResidualLayout<NDimSpatial>>, OutputLayout<NDimSpatial>,
......@@ -500,7 +500,7 @@ using DeviceConvFwdInstances = std::tuple<
4, 1, // MRepeat x NRepeat
S<1, 64, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 16, 16, true, // ABlockTransferThreadClusterLengths_AK0_M_AK1, ArrangeOrder, SrcAccessOrder, VectorDim, SrcScalarPerVector, DstScalarPerVector_AK1, LdsExtraM
S<1, 32, 2>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, true,
1, 1, S<1, 32, 1, 2>, 16>,
4, 1, S<1, 32, 1, 2>, 16>,
// K0 = 8
ck::tensor_operation::device::DeviceGroupedConvFwdMultipleD_Wmma_CShuffle<
NDimSpatial,
......@@ -512,7 +512,7 @@ using DeviceConvFwdInstances = std::tuple<
4, 1, // MRepeat x NRepeat
S<1, 256, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 16, 16, true, // ABlockTransferThreadClusterLengths_AK0_M_AK1, ArrangeOrder, SrcAccessOrder, VectorDim, SrcScalarPerVector, DstScalarPerVector_AK1, LdsExtraM
S<1, 32, 8>, S<1, 0, 2>, S<1, 0, 2>, 2, 2, 2, true,
1, 1, S<1, 128, 1, 2>, 16>,
4, 1, S<1, 128, 1, 2>, 16>,
ck::tensor_operation::device::DeviceGroupedConvFwdMultipleD_Wmma_CShuffle<
NDimSpatial,
InputLayout<NDimSpatial>, WeightLayout<NDimSpatial>, ck::Tuple<BiasLayout<NDimSpatial>, ResidualLayout<NDimSpatial>>, OutputLayout<NDimSpatial>,
......@@ -523,7 +523,7 @@ using DeviceConvFwdInstances = std::tuple<
4, 1, // MRepeat x NRepeat
S<1, 128, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 16, 16, true, // ABlockTransferThreadClusterLengths_AK0_M_AK1, ArrangeOrder, SrcAccessOrder, VectorDim, SrcScalarPerVector, DstScalarPerVector_AK1, LdsExtraM
S<1, 32, 4>, S<1, 0, 2>, S<1, 0, 2>, 2, 4, 4, true,
1, 1, S<1, 64, 1, 2>, 16>,
4, 1, S<1, 64, 1, 2>, 16>,
ck::tensor_operation::device::DeviceGroupedConvFwdMultipleD_Wmma_CShuffle<
NDimSpatial,
InputLayout<NDimSpatial>, WeightLayout<NDimSpatial>, ck::Tuple<BiasLayout<NDimSpatial>, ResidualLayout<NDimSpatial>>, OutputLayout<NDimSpatial>,
......@@ -534,7 +534,7 @@ using DeviceConvFwdInstances = std::tuple<
4, 1, // MRepeat x NRepeat
S<1, 64, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 16, 16, true, // ABlockTransferThreadClusterLengths_AK0_M_AK1, ArrangeOrder, SrcAccessOrder, VectorDim, SrcScalarPerVector, DstScalarPerVector_AK1, LdsExtraM
S<1, 32, 2>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, true,
1, 1, S<1, 32, 1, 2>, 16>,
4, 1, S<1, 32, 1, 2>, 16>,
// GEMM_N=32, Wave N = 1
// K0 = 2
......@@ -548,7 +548,7 @@ using DeviceConvFwdInstances = std::tuple<
2, 2, // MRepeat x NRepeat
S<1, 256, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 16, 16, true, // ABlockTransferThreadClusterLengths_AK0_M_AK1, ArrangeOrder, SrcAccessOrder, VectorDim, SrcScalarPerVector, DstScalarPerVector_AK1, LdsExtraM
S<1, 32, 8>, S<1, 0, 2>, S<1, 0, 2>, 2, 4, 4, true,
1, 1, S<1, 128, 1, 2>, 16>,
2, 2, S<1, 128, 1, 2>, 16>,
ck::tensor_operation::device::DeviceGroupedConvFwdMultipleD_Wmma_CShuffle<
NDimSpatial,
InputLayout<NDimSpatial>, WeightLayout<NDimSpatial>, ck::Tuple<BiasLayout<NDimSpatial>, ResidualLayout<NDimSpatial>>, OutputLayout<NDimSpatial>,
......@@ -559,7 +559,7 @@ using DeviceConvFwdInstances = std::tuple<
2, 2, // MRepeat x NRepeat
S<1, 128, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 16, 16, true, // ABlockTransferThreadClusterLengths_AK0_M_AK1, ArrangeOrder, SrcAccessOrder, VectorDim, SrcScalarPerVector, DstScalarPerVector_AK1, LdsExtraM
S<1, 32, 4>, S<1, 0, 2>, S<1, 0, 2>, 2, 4, 4, true,
1, 1, S<1, 64, 1, 2>, 16>,
2, 2, S<1, 64, 1, 2>, 16>,
ck::tensor_operation::device::DeviceGroupedConvFwdMultipleD_Wmma_CShuffle<
NDimSpatial,
InputLayout<NDimSpatial>, WeightLayout<NDimSpatial>, ck::Tuple<BiasLayout<NDimSpatial>, ResidualLayout<NDimSpatial>>, OutputLayout<NDimSpatial>,
......@@ -570,7 +570,7 @@ using DeviceConvFwdInstances = std::tuple<
2, 2, // MRepeat x NRepeat
S<1, 64, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 16, 16, true, // ABlockTransferThreadClusterLengths_AK0_M_AK1, ArrangeOrder, SrcAccessOrder, VectorDim, SrcScalarPerVector, DstScalarPerVector_AK1, LdsExtraM
S<1, 32, 2>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, true,
1, 1, S<1, 32, 1, 2>, 16>,
2, 2, S<1, 32, 1, 2>, 16>,
// K0 = 3
ck::tensor_operation::device::DeviceGroupedConvFwdMultipleD_Wmma_CShuffle<
NDimSpatial,
......@@ -582,7 +582,7 @@ using DeviceConvFwdInstances = std::tuple<
2, 2, // MRepeat x NRepeat
S<1, 256, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 16, 16, true, // ABlockTransferThreadClusterLengths_AK0_M_AK1, ArrangeOrder, SrcAccessOrder, VectorDim, SrcScalarPerVector, DstScalarPerVector_AK1, LdsExtraM
S<1, 32, 8>, S<1, 0, 2>, S<1, 0, 2>, 2, 4, 4, true,
1, 1, S<1, 128, 1, 2>, 16>,
2, 2, S<1, 128, 1, 2>, 16>,
ck::tensor_operation::device::DeviceGroupedConvFwdMultipleD_Wmma_CShuffle<
NDimSpatial,
InputLayout<NDimSpatial>, WeightLayout<NDimSpatial>, ck::Tuple<BiasLayout<NDimSpatial>, ResidualLayout<NDimSpatial>>, OutputLayout<NDimSpatial>,
......@@ -593,7 +593,7 @@ using DeviceConvFwdInstances = std::tuple<
2, 2, // MRepeat x NRepeat
S<1, 192, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 16, 16, true, // ABlockTransferThreadClusterLengths_AK0_M_AK1, ArrangeOrder, SrcAccessOrder, VectorDim, SrcScalarPerVector, DstScalarPerVector_AK1, LdsExtraM
S<3, 32, 2>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, true,
1, 1, S<1, 64, 1, 3>, 16>,
2, 2, S<1, 64, 1, 2>, 16>,
ck::tensor_operation::device::DeviceGroupedConvFwdMultipleD_Wmma_CShuffle<
NDimSpatial,
InputLayout<NDimSpatial>, WeightLayout<NDimSpatial>, ck::Tuple<BiasLayout<NDimSpatial>, ResidualLayout<NDimSpatial>>, OutputLayout<NDimSpatial>,
......@@ -604,7 +604,7 @@ using DeviceConvFwdInstances = std::tuple<
2, 2, // MRepeat x NRepeat
S<1, 128, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 16, 16, true, // ABlockTransferThreadClusterLengths_AK0_M_AK1, ArrangeOrder, SrcAccessOrder, VectorDim, SrcScalarPerVector, DstScalarPerVector_AK1, LdsExtraM
S<1, 32, 4>, S<1, 0, 2>, S<1, 0, 2>, 2, 4, 4, true,
1, 1, S<1, 64, 1, 2>, 16>,
2, 2, S<1, 64, 1, 2>, 16>,
ck::tensor_operation::device::DeviceGroupedConvFwdMultipleD_Wmma_CShuffle<
NDimSpatial,
InputLayout<NDimSpatial>, WeightLayout<NDimSpatial>, ck::Tuple<BiasLayout<NDimSpatial>, ResidualLayout<NDimSpatial>>, OutputLayout<NDimSpatial>,
......@@ -615,7 +615,7 @@ using DeviceConvFwdInstances = std::tuple<
2, 2, // MRepeat x NRepeat
S<1, 64, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 16, 16, true, // ABlockTransferThreadClusterLengths_AK0_M_AK1, ArrangeOrder, SrcAccessOrder, VectorDim, SrcScalarPerVector, DstScalarPerVector_AK1, LdsExtraM
S<1, 32, 2>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, true,
1, 1, S<1, 32, 1, 2>, 16>,
2, 2, S<1, 32, 1, 2>, 16>,
// K0 = 4
ck::tensor_operation::device::DeviceGroupedConvFwdMultipleD_Wmma_CShuffle<
NDimSpatial,
......@@ -627,7 +627,7 @@ using DeviceConvFwdInstances = std::tuple<
2, 2, // MRepeat x NRepeat
S<1, 256, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 16, 16, true, // ABlockTransferThreadClusterLengths_AK0_M_AK1, ArrangeOrder, SrcAccessOrder, VectorDim, SrcScalarPerVector, DstScalarPerVector_AK1, LdsExtraM
S<1, 32, 8>, S<1, 0, 2>, S<1, 0, 2>, 2, 2, 2, true,
1, 1, S<1, 128, 1, 2>, 16>,
2, 2, S<1, 128, 1, 2>, 16>,
ck::tensor_operation::device::DeviceGroupedConvFwdMultipleD_Wmma_CShuffle<
NDimSpatial,
InputLayout<NDimSpatial>, WeightLayout<NDimSpatial>, ck::Tuple<BiasLayout<NDimSpatial>, ResidualLayout<NDimSpatial>>, OutputLayout<NDimSpatial>,
......@@ -638,7 +638,7 @@ using DeviceConvFwdInstances = std::tuple<
2, 2, // MRepeat x NRepeat
S<1, 128, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 16, 16, true, // ABlockTransferThreadClusterLengths_AK0_M_AK1, ArrangeOrder, SrcAccessOrder, VectorDim, SrcScalarPerVector, DstScalarPerVector_AK1, LdsExtraM
S<1, 32, 4>, S<1, 0, 2>, S<1, 0, 2>, 2, 4, 4, true,
1, 1, S<1, 64, 1, 2>, 16>,
2, 2, S<1, 64, 1, 2>, 16>,
ck::tensor_operation::device::DeviceGroupedConvFwdMultipleD_Wmma_CShuffle<
NDimSpatial,
InputLayout<NDimSpatial>, WeightLayout<NDimSpatial>, ck::Tuple<BiasLayout<NDimSpatial>, ResidualLayout<NDimSpatial>>, OutputLayout<NDimSpatial>,
......@@ -649,7 +649,7 @@ using DeviceConvFwdInstances = std::tuple<
2, 2, // MRepeat x NRepeat
S<1, 64, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 16, 16, true, // ABlockTransferThreadClusterLengths_AK0_M_AK1, ArrangeOrder, SrcAccessOrder, VectorDim, SrcScalarPerVector, DstScalarPerVector_AK1, LdsExtraM
S<1, 32, 2>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, true,
1, 1, S<1, 32, 1, 2>, 16>,
2, 2, S<1, 32, 1, 2>, 16>,
// K0 = 5
ck::tensor_operation::device::DeviceGroupedConvFwdMultipleD_Wmma_CShuffle<
NDimSpatial,
......@@ -661,7 +661,7 @@ using DeviceConvFwdInstances = std::tuple<
2, 2, // MRepeat x NRepeat
S<1, 256, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 16, 16, true, // ABlockTransferThreadClusterLengths_AK0_M_AK1, ArrangeOrder, SrcAccessOrder, VectorDim, SrcScalarPerVector, DstScalarPerVector_AK1, LdsExtraM
S<1, 32, 8>, S<1, 0, 2>, S<1, 0, 2>, 2, 2, 2, true,
1, 1, S<1, 128, 1, 2>, 16>,
2, 2, S<1, 128, 1, 2>, 16>,
ck::tensor_operation::device::DeviceGroupedConvFwdMultipleD_Wmma_CShuffle<
NDimSpatial,
InputLayout<NDimSpatial>, WeightLayout<NDimSpatial>, ck::Tuple<BiasLayout<NDimSpatial>, ResidualLayout<NDimSpatial>>, OutputLayout<NDimSpatial>,
......@@ -672,7 +672,7 @@ using DeviceConvFwdInstances = std::tuple<
2, 2, // MRepeat x NRepeat
S<1, 128, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 16, 16, true, // ABlockTransferThreadClusterLengths_AK0_M_AK1, ArrangeOrder, SrcAccessOrder, VectorDim, SrcScalarPerVector, DstScalarPerVector_AK1, LdsExtraM
S<1, 32, 4>, S<1, 0, 2>, S<1, 0, 2>, 2, 4, 4, true,
1, 1, S<1, 64, 1, 2>, 16>,
2, 2, S<1, 64, 1, 2>, 16>,
ck::tensor_operation::device::DeviceGroupedConvFwdMultipleD_Wmma_CShuffle<
NDimSpatial,
InputLayout<NDimSpatial>, WeightLayout<NDimSpatial>, ck::Tuple<BiasLayout<NDimSpatial>, ResidualLayout<NDimSpatial>>, OutputLayout<NDimSpatial>,
......@@ -683,7 +683,7 @@ using DeviceConvFwdInstances = std::tuple<
2, 2, // MRepeat x NRepeat
S<1, 64, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 16, 16, true, // ABlockTransferThreadClusterLengths_AK0_M_AK1, ArrangeOrder, SrcAccessOrder, VectorDim, SrcScalarPerVector, DstScalarPerVector_AK1, LdsExtraM
S<1, 32, 2>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, true,
1, 1, S<1, 32, 1, 2>, 16>,
2, 2, S<1, 32, 1, 2>, 16>,
// K0 = 6
ck::tensor_operation::device::DeviceGroupedConvFwdMultipleD_Wmma_CShuffle<
NDimSpatial,
......@@ -695,7 +695,7 @@ using DeviceConvFwdInstances = std::tuple<
2, 2, // MRepeat x NRepeat
S<1, 256, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 16, 16, true, // ABlockTransferThreadClusterLengths_AK0_M_AK1, ArrangeOrder, SrcAccessOrder, VectorDim, SrcScalarPerVector, DstScalarPerVector_AK1, LdsExtraM
S<1, 32, 8>, S<1, 0, 2>, S<1, 0, 2>, 2, 2, 2, true,
1, 1, S<1, 128, 1, 2>, 16>,
2, 2, S<1, 128, 1, 2>, 16>,
ck::tensor_operation::device::DeviceGroupedConvFwdMultipleD_Wmma_CShuffle<
NDimSpatial,
InputLayout<NDimSpatial>, WeightLayout<NDimSpatial>, ck::Tuple<BiasLayout<NDimSpatial>, ResidualLayout<NDimSpatial>>, OutputLayout<NDimSpatial>,
......@@ -706,7 +706,7 @@ using DeviceConvFwdInstances = std::tuple<
2, 2, // MRepeat x NRepeat
S<1, 128, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 16, 16, true, // ABlockTransferThreadClusterLengths_AK0_M_AK1, ArrangeOrder, SrcAccessOrder, VectorDim, SrcScalarPerVector, DstScalarPerVector_AK1, LdsExtraM
S<1, 32, 4>, S<1, 0, 2>, S<1, 0, 2>, 2, 4, 4, true,
1, 1, S<1, 64, 1, 2>, 16>,
2, 2, S<1, 64, 1, 2>, 16>,
ck::tensor_operation::device::DeviceGroupedConvFwdMultipleD_Wmma_CShuffle<
NDimSpatial,
InputLayout<NDimSpatial>, WeightLayout<NDimSpatial>, ck::Tuple<BiasLayout<NDimSpatial>, ResidualLayout<NDimSpatial>>, OutputLayout<NDimSpatial>,
......@@ -717,7 +717,7 @@ using DeviceConvFwdInstances = std::tuple<
2, 2, // MRepeat x NRepeat
S<1, 64, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 16, 16, true, // ABlockTransferThreadClusterLengths_AK0_M_AK1, ArrangeOrder, SrcAccessOrder, VectorDim, SrcScalarPerVector, DstScalarPerVector_AK1, LdsExtraM
S<1, 32, 2>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, true,
1, 1, S<1, 32, 1, 2>, 16>,
2, 2, S<1, 32, 1, 2>, 16>,
// K0 = 7
ck::tensor_operation::device::DeviceGroupedConvFwdMultipleD_Wmma_CShuffle<
NDimSpatial,
......@@ -729,7 +729,7 @@ using DeviceConvFwdInstances = std::tuple<
2, 2, // MRepeat x NRepeat
S<1, 256, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 16, 16, true, // ABlockTransferThreadClusterLengths_AK0_M_AK1, ArrangeOrder, SrcAccessOrder, VectorDim, SrcScalarPerVector, DstScalarPerVector_AK1, LdsExtraM
S<1, 32, 8>, S<1, 0, 2>, S<1, 0, 2>, 2, 2, 2, true,
1, 1, S<1, 128, 1, 2>, 16>,
2, 2, S<1, 128, 1, 2>, 16>,
ck::tensor_operation::device::DeviceGroupedConvFwdMultipleD_Wmma_CShuffle<
NDimSpatial,
InputLayout<NDimSpatial>, WeightLayout<NDimSpatial>, ck::Tuple<BiasLayout<NDimSpatial>, ResidualLayout<NDimSpatial>>, OutputLayout<NDimSpatial>,
......@@ -740,7 +740,7 @@ using DeviceConvFwdInstances = std::tuple<
2, 2, // MRepeat x NRepeat
S<1, 128, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 16, 16, true, // ABlockTransferThreadClusterLengths_AK0_M_AK1, ArrangeOrder, SrcAccessOrder, VectorDim, SrcScalarPerVector, DstScalarPerVector_AK1, LdsExtraM
S<1, 32, 4>, S<1, 0, 2>, S<1, 0, 2>, 2, 4, 4, true,
1, 1, S<1, 64, 1, 2>, 16>,
2, 2, S<1, 64, 1, 2>, 16>,
ck::tensor_operation::device::DeviceGroupedConvFwdMultipleD_Wmma_CShuffle<
NDimSpatial,
InputLayout<NDimSpatial>, WeightLayout<NDimSpatial>, ck::Tuple<BiasLayout<NDimSpatial>, ResidualLayout<NDimSpatial>>, OutputLayout<NDimSpatial>,
......@@ -751,7 +751,7 @@ using DeviceConvFwdInstances = std::tuple<
2, 2, // MRepeat x NRepeat
S<1, 64, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 16, 16, true, // ABlockTransferThreadClusterLengths_AK0_M_AK1, ArrangeOrder, SrcAccessOrder, VectorDim, SrcScalarPerVector, DstScalarPerVector_AK1, LdsExtraM
S<1, 32, 2>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, true,
1, 1, S<1, 32, 1, 2>, 16>,
2, 2, S<1, 32, 1, 2>, 16>,
// K0 = 8
ck::tensor_operation::device::DeviceGroupedConvFwdMultipleD_Wmma_CShuffle<
NDimSpatial,
......@@ -763,7 +763,7 @@ using DeviceConvFwdInstances = std::tuple<
2, 2, // MRepeat x NRepeat
S<1, 256, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 16, 16, true, // ABlockTransferThreadClusterLengths_AK0_M_AK1, ArrangeOrder, SrcAccessOrder, VectorDim, SrcScalarPerVector, DstScalarPerVector_AK1, LdsExtraM
S<1, 32, 8>, S<1, 0, 2>, S<1, 0, 2>, 2, 2, 2, true,
1, 1, S<1, 128, 1, 2>, 16>,
2, 2, S<1, 128, 1, 2>, 16>,
ck::tensor_operation::device::DeviceGroupedConvFwdMultipleD_Wmma_CShuffle<
NDimSpatial,
InputLayout<NDimSpatial>, WeightLayout<NDimSpatial>, ck::Tuple<BiasLayout<NDimSpatial>, ResidualLayout<NDimSpatial>>, OutputLayout<NDimSpatial>,
......@@ -774,7 +774,7 @@ using DeviceConvFwdInstances = std::tuple<
2, 2, // MRepeat x NRepeat
S<1, 128, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 16, 16, true, // ABlockTransferThreadClusterLengths_AK0_M_AK1, ArrangeOrder, SrcAccessOrder, VectorDim, SrcScalarPerVector, DstScalarPerVector_AK1, LdsExtraM
S<1, 32, 4>, S<1, 0, 2>, S<1, 0, 2>, 2, 4, 4, true,
1, 1, S<1, 64, 1, 2>, 16>,
2, 2, S<1, 64, 1, 2>, 16>,
ck::tensor_operation::device::DeviceGroupedConvFwdMultipleD_Wmma_CShuffle<
NDimSpatial,
InputLayout<NDimSpatial>, WeightLayout<NDimSpatial>, ck::Tuple<BiasLayout<NDimSpatial>, ResidualLayout<NDimSpatial>>, OutputLayout<NDimSpatial>,
......@@ -785,7 +785,7 @@ using DeviceConvFwdInstances = std::tuple<
2, 2, // MRepeat x NRepeat
S<1, 64, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 16, 16, true, // ABlockTransferThreadClusterLengths_AK0_M_AK1, ArrangeOrder, SrcAccessOrder, VectorDim, SrcScalarPerVector, DstScalarPerVector_AK1, LdsExtraM
S<1, 32, 2>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, true,
1, 1, S<1, 32, 1, 2>, 16>,
2, 2, S<1, 32, 1, 2>, 16>,
// GEMM_N=64, Wave N = 1
// K0 = 2
ck::tensor_operation::device::DeviceGroupedConvFwdMultipleD_Wmma_CShuffle<
......@@ -798,7 +798,7 @@ using DeviceConvFwdInstances = std::tuple<
2, 4, // MRepeat x NRepeat
S<1, 256, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 16, 16, true, // ABlockTransferThreadClusterLengths_AK0_M_AK1, ArrangeOrder, SrcAccessOrder, VectorDim, SrcScalarPerVector, DstScalarPerVector_AK1, LdsExtraM
S<1, 64, 4>, S<1, 0, 2>, S<1, 0, 2>, 2, 4, 4, true,
1, 1, S<1, 64, 1, 4>, 16>,
2, 4, S<1, 64, 1, 4>, 16>,
ck::tensor_operation::device::DeviceGroupedConvFwdMultipleD_Wmma_CShuffle<
NDimSpatial,
InputLayout<NDimSpatial>, WeightLayout<NDimSpatial>, ck::Tuple<BiasLayout<NDimSpatial>, ResidualLayout<NDimSpatial>>, OutputLayout<NDimSpatial>,
......@@ -809,7 +809,7 @@ using DeviceConvFwdInstances = std::tuple<
2, 4, // MRepeat x NRepeat
S<1, 128, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 16, 16, true, // ABlockTransferThreadClusterLengths_AK0_M_AK1, ArrangeOrder, SrcAccessOrder, VectorDim, SrcScalarPerVector, DstScalarPerVector_AK1, LdsExtraM
S<1, 64, 2>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, true,
1, 1, S<1, 32, 1, 4>, 16>,
2, 4, S<1, 32, 1, 4>, 16>,
ck::tensor_operation::device::DeviceGroupedConvFwdMultipleD_Wmma_CShuffle<
NDimSpatial,
InputLayout<NDimSpatial>, WeightLayout<NDimSpatial>, ck::Tuple<BiasLayout<NDimSpatial>, ResidualLayout<NDimSpatial>>, OutputLayout<NDimSpatial>,
......@@ -820,7 +820,7 @@ using DeviceConvFwdInstances = std::tuple<
2, 4, // MRepeat x NRepeat
S<1, 64, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 16, 16, true, // ABlockTransferThreadClusterLengths_AK0_M_AK1, ArrangeOrder, SrcAccessOrder, VectorDim, SrcScalarPerVector, DstScalarPerVector_AK1, LdsExtraM
S<1, 64, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 16, 16, true,
1, 1, S<1, 16, 1, 4>, 16>,
2, 4, S<1, 16, 1, 4>, 16>,
// K0 = 3
ck::tensor_operation::device::DeviceGroupedConvFwdMultipleD_Wmma_CShuffle<
NDimSpatial,
......@@ -832,7 +832,7 @@ using DeviceConvFwdInstances = std::tuple<
2, 4, // MRepeat x NRepeat
S<1, 256, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 16, 16, true, // ABlockTransferThreadClusterLengths_AK0_M_AK1, ArrangeOrder, SrcAccessOrder, VectorDim, SrcScalarPerVector, DstScalarPerVector_AK1, LdsExtraM
S<1, 64, 4>, S<1, 0, 2>, S<1, 0, 2>, 2, 4, 4, true,
1, 1, S<1, 64, 1, 4>, 16>,
2, 4, S<1, 64, 1, 4>, 16>,
ck::tensor_operation::device::DeviceGroupedConvFwdMultipleD_Wmma_CShuffle<
NDimSpatial,
InputLayout<NDimSpatial>, WeightLayout<NDimSpatial>, ck::Tuple<BiasLayout<NDimSpatial>, ResidualLayout<NDimSpatial>>, OutputLayout<NDimSpatial>,
......@@ -843,7 +843,7 @@ using DeviceConvFwdInstances = std::tuple<
2, 4, // MRepeat x NRepeat
S<1, 192, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 16, 16, true, // ABlockTransferThreadClusterLengths_AK0_M_AK1, ArrangeOrder, SrcAccessOrder, VectorDim, SrcScalarPerVector, DstScalarPerVector_AK1, LdsExtraM
S<3, 64, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 16, 16, true,
1, 1, S<1, 64, 1, 3>, 16>,
2, 4, S<1, 48, 1, 4>, 16>,
ck::tensor_operation::device::DeviceGroupedConvFwdMultipleD_Wmma_CShuffle<
NDimSpatial,
InputLayout<NDimSpatial>, WeightLayout<NDimSpatial>, ck::Tuple<BiasLayout<NDimSpatial>, ResidualLayout<NDimSpatial>>, OutputLayout<NDimSpatial>,
......@@ -854,7 +854,7 @@ using DeviceConvFwdInstances = std::tuple<
2, 4, // MRepeat x NRepeat
S<1, 128, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 16, 16, true, // ABlockTransferThreadClusterLengths_AK0_M_AK1, ArrangeOrder, SrcAccessOrder, VectorDim, SrcScalarPerVector, DstScalarPerVector_AK1, LdsExtraM
S<1, 64, 2>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, true,
1, 1, S<1, 32, 1, 4>, 16>,
2, 4, S<1, 32, 1, 4>, 16>,
ck::tensor_operation::device::DeviceGroupedConvFwdMultipleD_Wmma_CShuffle<
NDimSpatial,
InputLayout<NDimSpatial>, WeightLayout<NDimSpatial>, ck::Tuple<BiasLayout<NDimSpatial>, ResidualLayout<NDimSpatial>>, OutputLayout<NDimSpatial>,
......@@ -865,7 +865,7 @@ using DeviceConvFwdInstances = std::tuple<
2, 4, // MRepeat x NRepeat
S<1, 64, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 16, 16, true, // ABlockTransferThreadClusterLengths_AK0_M_AK1, ArrangeOrder, SrcAccessOrder, VectorDim, SrcScalarPerVector, DstScalarPerVector_AK1, LdsExtraM
S<1, 64, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 16, 16, true,
1, 1, S<1, 16, 1, 4>, 16>,
2, 4, S<1, 16, 1, 4>, 16>,
// K0 = 4
ck::tensor_operation::device::DeviceGroupedConvFwdMultipleD_Wmma_CShuffle<
NDimSpatial,
......@@ -877,7 +877,7 @@ using DeviceConvFwdInstances = std::tuple<
2, 4, // MRepeat x NRepeat
S<1, 256, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 16, 16, true, // ABlockTransferThreadClusterLengths_AK0_M_AK1, ArrangeOrder, SrcAccessOrder, VectorDim, SrcScalarPerVector, DstScalarPerVector_AK1, LdsExtraM
S<1, 64, 4>, S<1, 0, 2>, S<1, 0, 2>, 2, 4, 4, true,
1, 1, S<1, 64, 1, 4>, 16>,
2, 4, S<1, 64, 1, 4>, 16>,
ck::tensor_operation::device::DeviceGroupedConvFwdMultipleD_Wmma_CShuffle<
NDimSpatial,
InputLayout<NDimSpatial>, WeightLayout<NDimSpatial>, ck::Tuple<BiasLayout<NDimSpatial>, ResidualLayout<NDimSpatial>>, OutputLayout<NDimSpatial>,
......@@ -888,7 +888,7 @@ using DeviceConvFwdInstances = std::tuple<
2, 4, // MRepeat x NRepeat
S<1, 128, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 16, 16, true, // ABlockTransferThreadClusterLengths_AK0_M_AK1, ArrangeOrder, SrcAccessOrder, VectorDim, SrcScalarPerVector, DstScalarPerVector_AK1, LdsExtraM
S<1, 64, 2>, S<1, 0, 2>, S<1, 0, 2>, 2, 8 , 8, true,
1, 1, S<1, 32, 1, 4>, 16>,
2, 4, S<1, 32, 1, 4>, 16>,
ck::tensor_operation::device::DeviceGroupedConvFwdMultipleD_Wmma_CShuffle<
NDimSpatial,
InputLayout<NDimSpatial>, WeightLayout<NDimSpatial>, ck::Tuple<BiasLayout<NDimSpatial>, ResidualLayout<NDimSpatial>>, OutputLayout<NDimSpatial>,
......@@ -899,7 +899,7 @@ using DeviceConvFwdInstances = std::tuple<
2, 4, // MRepeat x NRepeat
S<1, 64, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 16, 16, true, // ABlockTransferThreadClusterLengths_AK0_M_AK1, ArrangeOrder, SrcAccessOrder, VectorDim, SrcScalarPerVector, DstScalarPerVector_AK1, LdsExtraM
S<1, 64, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 16, 16, true,
1, 1, S<1, 16, 1, 4>, 16>,
2, 4, S<1, 16, 1, 4>, 16>,
// K0 = 5
ck::tensor_operation::device::DeviceGroupedConvFwdMultipleD_Wmma_CShuffle<
NDimSpatial,
......@@ -911,7 +911,7 @@ using DeviceConvFwdInstances = std::tuple<
2, 4, // MRepeat x NRepeat
S<1, 256, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 16, 16, true, // ABlockTransferThreadClusterLengths_AK0_M_AK1, ArrangeOrder, SrcAccessOrder, VectorDim, SrcScalarPerVector, DstScalarPerVector_AK1, LdsExtraM
S<1, 64, 4>, S<1, 0, 2>, S<1, 0, 2>, 2, 4, 4, true,
1, 1, S<1, 64, 1, 4>, 16>,
2, 4, S<1, 64, 1, 4>, 16>,
ck::tensor_operation::device::DeviceGroupedConvFwdMultipleD_Wmma_CShuffle<
NDimSpatial,
InputLayout<NDimSpatial>, WeightLayout<NDimSpatial>, ck::Tuple<BiasLayout<NDimSpatial>, ResidualLayout<NDimSpatial>>, OutputLayout<NDimSpatial>,
......@@ -922,7 +922,7 @@ using DeviceConvFwdInstances = std::tuple<
2, 4, // MRepeat x NRepeat
S<1, 128, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 16, 16, true, // ABlockTransferThreadClusterLengths_AK0_M_AK1, ArrangeOrder, SrcAccessOrder, VectorDim, SrcScalarPerVector, DstScalarPerVector_AK1, LdsExtraM
S<1, 64, 2>, S<1, 0, 2>, S<1, 0, 2>, 2, 8 , 8, true,
1, 1, S<1, 32, 1, 4>, 16>,
2, 4, S<1, 32, 1, 4>, 16>,
ck::tensor_operation::device::DeviceGroupedConvFwdMultipleD_Wmma_CShuffle<
NDimSpatial,
InputLayout<NDimSpatial>, WeightLayout<NDimSpatial>, ck::Tuple<BiasLayout<NDimSpatial>, ResidualLayout<NDimSpatial>>, OutputLayout<NDimSpatial>,
......@@ -933,7 +933,7 @@ using DeviceConvFwdInstances = std::tuple<
2, 4, // MRepeat x NRepeat
S<1, 64, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 16, 16, true, // ABlockTransferThreadClusterLengths_AK0_M_AK1, ArrangeOrder, SrcAccessOrder, VectorDim, SrcScalarPerVector, DstScalarPerVector_AK1, LdsExtraM
S<1, 64, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 16, 16, true,
1, 1, S<1, 16, 1, 4>, 16>,
2, 4, S<1, 16, 1, 4>, 16>,
// K0 = 6
ck::tensor_operation::device::DeviceGroupedConvFwdMultipleD_Wmma_CShuffle<
NDimSpatial,
......@@ -945,7 +945,7 @@ using DeviceConvFwdInstances = std::tuple<
2, 4, // MRepeat x NRepeat
S<1, 256, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 16, 16, true, // ABlockTransferThreadClusterLengths_AK0_M_AK1, ArrangeOrder, SrcAccessOrder, VectorDim, SrcScalarPerVector, DstScalarPerVector_AK1, LdsExtraM
S<1, 64, 4>, S<1, 0, 2>, S<1, 0, 2>, 2, 4, 4, true,
1, 1, S<1, 64, 1, 4>, 16>,
2, 4, S<1, 64, 1, 4>, 16>,
ck::tensor_operation::device::DeviceGroupedConvFwdMultipleD_Wmma_CShuffle<
NDimSpatial,
InputLayout<NDimSpatial>, WeightLayout<NDimSpatial>, ck::Tuple<BiasLayout<NDimSpatial>, ResidualLayout<NDimSpatial>>, OutputLayout<NDimSpatial>,
......@@ -956,7 +956,7 @@ using DeviceConvFwdInstances = std::tuple<
2, 4, // MRepeat x NRepeat
S<1, 128, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 16, 16, true, // ABlockTransferThreadClusterLengths_AK0_M_AK1, ArrangeOrder, SrcAccessOrder, VectorDim, SrcScalarPerVector, DstScalarPerVector_AK1, LdsExtraM
S<1, 64, 2>, S<1, 0, 2>, S<1, 0, 2>, 2, 8 , 8, true,
1, 1, S<1, 32, 1, 4>, 16>,
2, 4, S<1, 32, 1, 4>, 16>,
ck::tensor_operation::device::DeviceGroupedConvFwdMultipleD_Wmma_CShuffle<
NDimSpatial,
InputLayout<NDimSpatial>, WeightLayout<NDimSpatial>, ck::Tuple<BiasLayout<NDimSpatial>, ResidualLayout<NDimSpatial>>, OutputLayout<NDimSpatial>,
......@@ -967,7 +967,7 @@ using DeviceConvFwdInstances = std::tuple<
2, 4, // MRepeat x NRepeat
S<1, 64, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 16, 16, true, // ABlockTransferThreadClusterLengths_AK0_M_AK1, ArrangeOrder, SrcAccessOrder, VectorDim, SrcScalarPerVector, DstScalarPerVector_AK1, LdsExtraM
S<1, 64, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 16, 16, true,
1, 1, S<1, 16, 1, 4>, 16>,
2, 4, S<1, 16, 1, 4>, 16>,
// K0 = 7
ck::tensor_operation::device::DeviceGroupedConvFwdMultipleD_Wmma_CShuffle<
NDimSpatial,
......@@ -979,7 +979,7 @@ using DeviceConvFwdInstances = std::tuple<
2, 4, // MRepeat x NRepeat
S<1, 256, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 16, 16, true, // ABlockTransferThreadClusterLengths_AK0_M_AK1, ArrangeOrder, SrcAccessOrder, VectorDim, SrcScalarPerVector, DstScalarPerVector_AK1, LdsExtraM
S<1, 64, 4>, S<1, 0, 2>, S<1, 0, 2>, 2, 4, 4, true,
1, 1, S<1, 64, 1, 4>, 16>,
2, 4, S<1, 64, 1, 4>, 16>,
ck::tensor_operation::device::DeviceGroupedConvFwdMultipleD_Wmma_CShuffle<
NDimSpatial,
InputLayout<NDimSpatial>, WeightLayout<NDimSpatial>, ck::Tuple<BiasLayout<NDimSpatial>, ResidualLayout<NDimSpatial>>, OutputLayout<NDimSpatial>,
......@@ -990,7 +990,7 @@ using DeviceConvFwdInstances = std::tuple<
2, 4, // MRepeat x NRepeat
S<1, 128, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 16, 16, true, // ABlockTransferThreadClusterLengths_AK0_M_AK1, ArrangeOrder, SrcAccessOrder, VectorDim, SrcScalarPerVector, DstScalarPerVector_AK1, LdsExtraM
S<1, 64, 2>, S<1, 0, 2>, S<1, 0, 2>, 2, 8 , 8, true,
1, 1, S<1, 32, 1, 4>, 16>,
2, 4, S<1, 32, 1, 4>, 16>,
ck::tensor_operation::device::DeviceGroupedConvFwdMultipleD_Wmma_CShuffle<
NDimSpatial,
InputLayout<NDimSpatial>, WeightLayout<NDimSpatial>, ck::Tuple<BiasLayout<NDimSpatial>, ResidualLayout<NDimSpatial>>, OutputLayout<NDimSpatial>,
......@@ -1001,7 +1001,7 @@ using DeviceConvFwdInstances = std::tuple<
2, 4, // MRepeat x NRepeat
S<1, 64, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 16, 16, true, // ABlockTransferThreadClusterLengths_AK0_M_AK1, ArrangeOrder, SrcAccessOrder, VectorDim, SrcScalarPerVector, DstScalarPerVector_AK1, LdsExtraM
S<1, 64, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 16, 16, true,
1, 1, S<1, 16, 1, 4>, 16>,
2, 4, S<1, 16, 1, 4>, 16>,
// K0 = 8
ck::tensor_operation::device::DeviceGroupedConvFwdMultipleD_Wmma_CShuffle<
NDimSpatial,
......@@ -1013,7 +1013,7 @@ using DeviceConvFwdInstances = std::tuple<
2, 4, // MRepeat x NRepeat
S<1, 256, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 16, 16, true, // ABlockTransferThreadClusterLengths_AK0_M_AK1, ArrangeOrder, SrcAccessOrder, VectorDim, SrcScalarPerVector, DstScalarPerVector_AK1, LdsExtraM
S<1, 64, 4>, S<1, 0, 2>, S<1, 0, 2>, 2, 4, 4, true,
1, 1, S<1, 64, 1, 4>, 16>,
2, 4, S<1, 64, 1, 4>, 16>,
ck::tensor_operation::device::DeviceGroupedConvFwdMultipleD_Wmma_CShuffle<
NDimSpatial,
InputLayout<NDimSpatial>, WeightLayout<NDimSpatial>, ck::Tuple<BiasLayout<NDimSpatial>, ResidualLayout<NDimSpatial>>, OutputLayout<NDimSpatial>,
......@@ -1024,7 +1024,7 @@ using DeviceConvFwdInstances = std::tuple<
2, 4, // MRepeat x NRepeat
S<1, 128, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 16, 16, true, // ABlockTransferThreadClusterLengths_AK0_M_AK1, ArrangeOrder, SrcAccessOrder, VectorDim, SrcScalarPerVector, DstScalarPerVector_AK1, LdsExtraM
S<1, 64, 2>, S<1, 0, 2>, S<1, 0, 2>, 2, 8 , 8, true,
1, 1, S<1, 32, 1, 4>, 16>,
2, 4, S<1, 32, 1, 4>, 16>,
ck::tensor_operation::device::DeviceGroupedConvFwdMultipleD_Wmma_CShuffle<
NDimSpatial,
InputLayout<NDimSpatial>, WeightLayout<NDimSpatial>, ck::Tuple<BiasLayout<NDimSpatial>, ResidualLayout<NDimSpatial>>, OutputLayout<NDimSpatial>,
......@@ -1035,7 +1035,7 @@ using DeviceConvFwdInstances = std::tuple<
2, 4, // MRepeat x NRepeat
S<1, 64, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 16, 16, true, // ABlockTransferThreadClusterLengths_AK0_M_AK1, ArrangeOrder, SrcAccessOrder, VectorDim, SrcScalarPerVector, DstScalarPerVector_AK1, LdsExtraM
S<1, 64, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 16, 16, true,
1, 1, S<1, 16, 1, 4>, 16>,
2, 4, S<1, 16, 1, 4>, 16>,
// GEMM_N = 64, WaveN = 2
ck::tensor_operation::device::DeviceGroupedConvFwdMultipleD_Wmma_CShuffle<
......@@ -1048,7 +1048,7 @@ using DeviceConvFwdInstances = std::tuple<
4, 2, // MRepeat x NRepeat
S<1, 256, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 16, 16, true, // ABlockTransferThreadClusterLengths_AK0_M_AK1, ArrangeOrder, SrcAccessOrder, VectorDim, SrcScalarPerVector, DstScalarPerVector_AK1, LdsExtraM
S<1, 64, 4>, S<1, 0, 2>, S<1, 0, 2>, 2, 4, 4, true,
1, 1, S<1, 64, 1, 4>, 16>,
4, 2, S<1, 64, 1, 4>, 16>,
ck::tensor_operation::device::DeviceGroupedConvFwdMultipleD_Wmma_CShuffle<
NDimSpatial,
InputLayout<NDimSpatial>, WeightLayout<NDimSpatial>, ck::Tuple<BiasLayout<NDimSpatial>, ResidualLayout<NDimSpatial>>, OutputLayout<NDimSpatial>,
......@@ -1059,7 +1059,7 @@ using DeviceConvFwdInstances = std::tuple<
4, 2, // MRepeat x NRepeat
S<1, 128, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 16, 16, true, // ABlockTransferThreadClusterLengths_AK0_M_AK1, ArrangeOrder, SrcAccessOrder, VectorDim, SrcScalarPerVector, DstScalarPerVector_AK1, LdsExtraM
S<1, 64, 2>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, true,
1, 1, S<1, 32, 1, 4>, 16>,
4, 2, S<1, 32, 1, 4>, 16>,
ck::tensor_operation::device::DeviceGroupedConvFwdMultipleD_Wmma_CShuffle<
NDimSpatial,
InputLayout<NDimSpatial>, WeightLayout<NDimSpatial>, ck::Tuple<BiasLayout<NDimSpatial>, ResidualLayout<NDimSpatial>>, OutputLayout<NDimSpatial>,
......@@ -1070,7 +1070,7 @@ using DeviceConvFwdInstances = std::tuple<
4, 2, // MRepeat x NRepeat
S<1, 64, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 16, 16, true, // ABlockTransferThreadClusterLengths_AK0_M_AK1, ArrangeOrder, SrcAccessOrder, VectorDim, SrcScalarPerVector, DstScalarPerVector_AK1, LdsExtraM
S<1, 64, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 16, 16, true,
1, 1, S<1, 16, 1, 4>, 16>,
4, 2, S<1, 16, 1, 4>, 16>,
// K0 = 3
ck::tensor_operation::device::DeviceGroupedConvFwdMultipleD_Wmma_CShuffle<
NDimSpatial,
......@@ -1082,7 +1082,7 @@ using DeviceConvFwdInstances = std::tuple<
4, 2, // MRepeat x NRepeat
S<1, 256, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 16, 16, true, // ABlockTransferThreadClusterLengths_AK0_M_AK1, ArrangeOrder, SrcAccessOrder, VectorDim, SrcScalarPerVector, DstScalarPerVector_AK1, LdsExtraM
S<1, 64, 4>, S<1, 0, 2>, S<1, 0, 2>, 2, 4, 4, true,
1, 1, S<1, 64, 1, 4>, 16>,
4, 2, S<1, 64, 1, 4>, 16>,
ck::tensor_operation::device::DeviceGroupedConvFwdMultipleD_Wmma_CShuffle<
NDimSpatial,
InputLayout<NDimSpatial>, WeightLayout<NDimSpatial>, ck::Tuple<BiasLayout<NDimSpatial>, ResidualLayout<NDimSpatial>>, OutputLayout<NDimSpatial>,
......@@ -1093,7 +1093,7 @@ using DeviceConvFwdInstances = std::tuple<
4, 2, // MRepeat x NRepeat
S<1, 192, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 16, 16, true, // ABlockTransferThreadClusterLengths_AK0_M_AK1, ArrangeOrder, SrcAccessOrder, VectorDim, SrcScalarPerVector, DstScalarPerVector_AK1, LdsExtraM
S<3, 64, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 16, 16, true,
1, 1, S<1, 64, 1, 3>, 16>,
4, 2, S<1, 48, 1, 4>, 16>,
ck::tensor_operation::device::DeviceGroupedConvFwdMultipleD_Wmma_CShuffle<
NDimSpatial,
InputLayout<NDimSpatial>, WeightLayout<NDimSpatial>, ck::Tuple<BiasLayout<NDimSpatial>, ResidualLayout<NDimSpatial>>, OutputLayout<NDimSpatial>,
......@@ -1104,7 +1104,7 @@ using DeviceConvFwdInstances = std::tuple<
4, 2, // MRepeat x NRepeat
S<1, 128, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 16, 16, true, // ABlockTransferThreadClusterLengths_AK0_M_AK1, ArrangeOrder, SrcAccessOrder, VectorDim, SrcScalarPerVector, DstScalarPerVector_AK1, LdsExtraM
S<1, 64, 2>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, true,
1, 1, S<1, 32, 1, 4>, 16>,
4, 2, S<1, 32, 1, 4>, 16>,
ck::tensor_operation::device::DeviceGroupedConvFwdMultipleD_Wmma_CShuffle<
NDimSpatial,
InputLayout<NDimSpatial>, WeightLayout<NDimSpatial>, ck::Tuple<BiasLayout<NDimSpatial>, ResidualLayout<NDimSpatial>>, OutputLayout<NDimSpatial>,
......@@ -1115,7 +1115,7 @@ using DeviceConvFwdInstances = std::tuple<
4, 2, // MRepeat x NRepeat
S<1, 64, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 16, 16, true, // ABlockTransferThreadClusterLengths_AK0_M_AK1, ArrangeOrder, SrcAccessOrder, VectorDim, SrcScalarPerVector, DstScalarPerVector_AK1, LdsExtraM
S<1, 64, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 16, 16, true,
1, 1, S<1, 16, 1, 4>, 16>,
4, 2, S<1, 16, 1, 4>, 16>,
// K0 = 4
ck::tensor_operation::device::DeviceGroupedConvFwdMultipleD_Wmma_CShuffle<
NDimSpatial,
......@@ -1127,7 +1127,7 @@ using DeviceConvFwdInstances = std::tuple<
4, 2, // MRepeat x NRepeat
S<1, 256, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 16, 16, true, // ABlockTransferThreadClusterLengths_AK0_M_AK1, ArrangeOrder, SrcAccessOrder, VectorDim, SrcScalarPerVector, DstScalarPerVector_AK1, LdsExtraM
S<1, 64, 4>, S<1, 0, 2>, S<1, 0, 2>, 2, 4, 4, true,
1, 1, S<1, 64, 1, 4>, 16>,
4, 2, S<1, 64, 1, 4>, 16>,
ck::tensor_operation::device::DeviceGroupedConvFwdMultipleD_Wmma_CShuffle<
NDimSpatial,
InputLayout<NDimSpatial>, WeightLayout<NDimSpatial>, ck::Tuple<BiasLayout<NDimSpatial>, ResidualLayout<NDimSpatial>>, OutputLayout<NDimSpatial>,
......@@ -1138,7 +1138,7 @@ using DeviceConvFwdInstances = std::tuple<
4, 2, // MRepeat x NRepeat
S<1, 128, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 16, 16, true, // ABlockTransferThreadClusterLengths_AK0_M_AK1, ArrangeOrder, SrcAccessOrder, VectorDim, SrcScalarPerVector, DstScalarPerVector_AK1, LdsExtraM
S<1, 64, 2>, S<1, 0, 2>, S<1, 0, 2>, 2, 8 , 8, true,
1, 1, S<1, 32, 1, 4>, 16>,
4, 2, S<1, 32, 1, 4>, 16>,
ck::tensor_operation::device::DeviceGroupedConvFwdMultipleD_Wmma_CShuffle<
NDimSpatial,
InputLayout<NDimSpatial>, WeightLayout<NDimSpatial>, ck::Tuple<BiasLayout<NDimSpatial>, ResidualLayout<NDimSpatial>>, OutputLayout<NDimSpatial>,
......@@ -1149,7 +1149,7 @@ using DeviceConvFwdInstances = std::tuple<
4, 2, // MRepeat x NRepeat
S<1, 64, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 16, 16, true, // ABlockTransferThreadClusterLengths_AK0_M_AK1, ArrangeOrder, SrcAccessOrder, VectorDim, SrcScalarPerVector, DstScalarPerVector_AK1, LdsExtraM
S<1, 64, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 16, 16, true,
1, 1, S<1, 16, 1, 4>, 16>,
4, 2, S<1, 16, 1, 4>, 16>,
// K0 = 5
ck::tensor_operation::device::DeviceGroupedConvFwdMultipleD_Wmma_CShuffle<
NDimSpatial,
......@@ -1161,7 +1161,7 @@ using DeviceConvFwdInstances = std::tuple<
4, 2, // MRepeat x NRepeat
S<1, 256, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 16, 16, true, // ABlockTransferThreadClusterLengths_AK0_M_AK1, ArrangeOrder, SrcAccessOrder, VectorDim, SrcScalarPerVector, DstScalarPerVector_AK1, LdsExtraM
S<1, 64, 4>, S<1, 0, 2>, S<1, 0, 2>, 2, 4, 4, true,
1, 1, S<1, 64, 1, 4>, 16>,
4, 2, S<1, 64, 1, 4>, 16>,
ck::tensor_operation::device::DeviceGroupedConvFwdMultipleD_Wmma_CShuffle<
NDimSpatial,
InputLayout<NDimSpatial>, WeightLayout<NDimSpatial>, ck::Tuple<BiasLayout<NDimSpatial>, ResidualLayout<NDimSpatial>>, OutputLayout<NDimSpatial>,
......@@ -1172,7 +1172,7 @@ using DeviceConvFwdInstances = std::tuple<
4, 2, // MRepeat x NRepeat
S<1, 128, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 16, 16, true, // ABlockTransferThreadClusterLengths_AK0_M_AK1, ArrangeOrder, SrcAccessOrder, VectorDim, SrcScalarPerVector, DstScalarPerVector_AK1, LdsExtraM
S<1, 64, 2>, S<1, 0, 2>, S<1, 0, 2>, 2, 8 , 8, true,
1, 1, S<1, 32, 1, 4>, 16>,
4, 2, S<1, 32, 1, 4>, 16>,
ck::tensor_operation::device::DeviceGroupedConvFwdMultipleD_Wmma_CShuffle<
NDimSpatial,
InputLayout<NDimSpatial>, WeightLayout<NDimSpatial>, ck::Tuple<BiasLayout<NDimSpatial>, ResidualLayout<NDimSpatial>>, OutputLayout<NDimSpatial>,
......@@ -1183,7 +1183,7 @@ using DeviceConvFwdInstances = std::tuple<
4, 2, // MRepeat x NRepeat
S<1, 64, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 16, 16, true, // ABlockTransferThreadClusterLengths_AK0_M_AK1, ArrangeOrder, SrcAccessOrder, VectorDim, SrcScalarPerVector, DstScalarPerVector_AK1, LdsExtraM
S<1, 64, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 16, 16, true,
1, 1, S<1, 16, 1, 4>, 16>,
4, 2, S<1, 16, 1, 4>, 16>,
// K0 = 6
ck::tensor_operation::device::DeviceGroupedConvFwdMultipleD_Wmma_CShuffle<
NDimSpatial,
......@@ -1195,7 +1195,7 @@ using DeviceConvFwdInstances = std::tuple<
4, 2, // MRepeat x NRepeat
S<1, 256, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 16, 16, true, // ABlockTransferThreadClusterLengths_AK0_M_AK1, ArrangeOrder, SrcAccessOrder, VectorDim, SrcScalarPerVector, DstScalarPerVector_AK1, LdsExtraM
S<1, 64, 4>, S<1, 0, 2>, S<1, 0, 2>, 2, 4, 4, true,
1, 1, S<1, 64, 1, 4>, 16>,
4, 2, S<1, 64, 1, 4>, 16>,
ck::tensor_operation::device::DeviceGroupedConvFwdMultipleD_Wmma_CShuffle<
NDimSpatial,
InputLayout<NDimSpatial>, WeightLayout<NDimSpatial>, ck::Tuple<BiasLayout<NDimSpatial>, ResidualLayout<NDimSpatial>>, OutputLayout<NDimSpatial>,
......@@ -1206,7 +1206,7 @@ using DeviceConvFwdInstances = std::tuple<
4, 2, // MRepeat x NRepeat
S<1, 128, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 16, 16, true, // ABlockTransferThreadClusterLengths_AK0_M_AK1, ArrangeOrder, SrcAccessOrder, VectorDim, SrcScalarPerVector, DstScalarPerVector_AK1, LdsExtraM
S<1, 64, 2>, S<1, 0, 2>, S<1, 0, 2>, 2, 8 , 8, true,
1, 1, S<1, 32, 1, 4>, 16>,
4, 2, S<1, 32, 1, 4>, 16>,
ck::tensor_operation::device::DeviceGroupedConvFwdMultipleD_Wmma_CShuffle<
NDimSpatial,
InputLayout<NDimSpatial>, WeightLayout<NDimSpatial>, ck::Tuple<BiasLayout<NDimSpatial>, ResidualLayout<NDimSpatial>>, OutputLayout<NDimSpatial>,
......@@ -1217,7 +1217,7 @@ using DeviceConvFwdInstances = std::tuple<
4, 2, // MRepeat x NRepeat
S<1, 64, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 16, 16, true, // ABlockTransferThreadClusterLengths_AK0_M_AK1, ArrangeOrder, SrcAccessOrder, VectorDim, SrcScalarPerVector, DstScalarPerVector_AK1, LdsExtraM
S<1, 64, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 16, 16, true,
1, 1, S<1, 16, 1, 4>, 16>,
4, 2, S<1, 16, 1, 4>, 16>,
// K0 = 7
ck::tensor_operation::device::DeviceGroupedConvFwdMultipleD_Wmma_CShuffle<
NDimSpatial,
......@@ -1229,7 +1229,7 @@ using DeviceConvFwdInstances = std::tuple<
4, 2, // MRepeat x NRepeat
S<1, 256, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 16, 16, true, // ABlockTransferThreadClusterLengths_AK0_M_AK1, ArrangeOrder, SrcAccessOrder, VectorDim, SrcScalarPerVector, DstScalarPerVector_AK1, LdsExtraM
S<1, 64, 4>, S<1, 0, 2>, S<1, 0, 2>, 2, 4, 4, true,
1, 1, S<1, 64, 1, 4>, 16>,
4, 2, S<1, 64, 1, 4>, 16>,
ck::tensor_operation::device::DeviceGroupedConvFwdMultipleD_Wmma_CShuffle<
NDimSpatial,
InputLayout<NDimSpatial>, WeightLayout<NDimSpatial>, ck::Tuple<BiasLayout<NDimSpatial>, ResidualLayout<NDimSpatial>>, OutputLayout<NDimSpatial>,
......@@ -1240,7 +1240,7 @@ using DeviceConvFwdInstances = std::tuple<
4, 2, // MRepeat x NRepeat
S<1, 128, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 16, 16, true, // ABlockTransferThreadClusterLengths_AK0_M_AK1, ArrangeOrder, SrcAccessOrder, VectorDim, SrcScalarPerVector, DstScalarPerVector_AK1, LdsExtraM
S<1, 64, 2>, S<1, 0, 2>, S<1, 0, 2>, 2, 8 , 8, true,
1, 1, S<1, 32, 1, 4>, 16>,
4, 2, S<1, 32, 1, 4>, 16>,
ck::tensor_operation::device::DeviceGroupedConvFwdMultipleD_Wmma_CShuffle<
NDimSpatial,
InputLayout<NDimSpatial>, WeightLayout<NDimSpatial>, ck::Tuple<BiasLayout<NDimSpatial>, ResidualLayout<NDimSpatial>>, OutputLayout<NDimSpatial>,
......@@ -1251,7 +1251,7 @@ using DeviceConvFwdInstances = std::tuple<
4, 2, // MRepeat x NRepeat
S<1, 64, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 16, 16, true, // ABlockTransferThreadClusterLengths_AK0_M_AK1, ArrangeOrder, SrcAccessOrder, VectorDim, SrcScalarPerVector, DstScalarPerVector_AK1, LdsExtraM
S<1, 64, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 16, 16, true,
1, 1, S<1, 16, 1, 4>, 16>,
4, 2, S<1, 16, 1, 4>, 16>,
// K0 = 8
ck::tensor_operation::device::DeviceGroupedConvFwdMultipleD_Wmma_CShuffle<
NDimSpatial,
......@@ -1263,7 +1263,7 @@ using DeviceConvFwdInstances = std::tuple<
4, 2, // MRepeat x NRepeat
S<1, 256, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 16, 16, true, // ABlockTransferThreadClusterLengths_AK0_M_AK1, ArrangeOrder, SrcAccessOrder, VectorDim, SrcScalarPerVector, DstScalarPerVector_AK1, LdsExtraM
S<1, 64, 4>, S<1, 0, 2>, S<1, 0, 2>, 2, 4, 4, true,
1, 1, S<1, 64, 1, 4>, 16>,
4, 2, S<1, 64, 1, 4>, 16>,
ck::tensor_operation::device::DeviceGroupedConvFwdMultipleD_Wmma_CShuffle<
NDimSpatial,
InputLayout<NDimSpatial>, WeightLayout<NDimSpatial>, ck::Tuple<BiasLayout<NDimSpatial>, ResidualLayout<NDimSpatial>>, OutputLayout<NDimSpatial>,
......@@ -1274,7 +1274,7 @@ using DeviceConvFwdInstances = std::tuple<
4, 2, // MRepeat x NRepeat
S<1, 128, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 16, 16, true, // ABlockTransferThreadClusterLengths_AK0_M_AK1, ArrangeOrder, SrcAccessOrder, VectorDim, SrcScalarPerVector, DstScalarPerVector_AK1, LdsExtraM
S<1, 64, 2>, S<1, 0, 2>, S<1, 0, 2>, 2, 8 , 8, true,
1, 1, S<1, 32, 1, 4>, 16>,
4, 2, S<1, 32, 1, 4>, 16>,
ck::tensor_operation::device::DeviceGroupedConvFwdMultipleD_Wmma_CShuffle<
NDimSpatial,
InputLayout<NDimSpatial>, WeightLayout<NDimSpatial>, ck::Tuple<BiasLayout<NDimSpatial>, ResidualLayout<NDimSpatial>>, OutputLayout<NDimSpatial>,
......@@ -1285,7 +1285,8 @@ using DeviceConvFwdInstances = std::tuple<
4, 2, // MRepeat x NRepeat
S<1, 64, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 16, 16, true, // ABlockTransferThreadClusterLengths_AK0_M_AK1, ArrangeOrder, SrcAccessOrder, VectorDim, SrcScalarPerVector, DstScalarPerVector_AK1, LdsExtraM
S<1, 64, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 16, 16, true,
1, 1, S<1, 16, 1, 4>, 16>,
4, 2, S<1, 16, 1, 4>, 16>,
// GEMM_N=48
ck::tensor_operation::device::DeviceGroupedConvFwdMultipleD_Wmma_CShuffle<
NDimSpatial,
InputLayout<NDimSpatial>, WeightLayout<NDimSpatial>, ck::Tuple<BiasLayout<NDimSpatial>, ResidualLayout<NDimSpatial>>, OutputLayout<NDimSpatial>,
......@@ -1296,7 +1297,7 @@ using DeviceConvFwdInstances = std::tuple<
6, 1, // MRepeat x NRepeat
S<1, 192, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 16, 16, true, // ABlockTransferThreadClusterLengths_AK0_M_AK1, ArrangeOrder, SrcAccessOrder, VectorDim, SrcScalarPerVector, DstScalarPerVector_AK1, LdsExtraM
S<1, 48, 4>, S<1, 0, 2>, S<1, 0, 2>, 2, 4, 4, true,
1, 1, S<1, 64, 1, 3>, 16>,
6, 1, S<1, 64, 1, 3>, 16>,
ck::tensor_operation::device::DeviceGroupedConvFwdMultipleD_Wmma_CShuffle<
NDimSpatial,
InputLayout<NDimSpatial>, WeightLayout<NDimSpatial>, ck::Tuple<BiasLayout<NDimSpatial>, ResidualLayout<NDimSpatial>>, OutputLayout<NDimSpatial>,
......@@ -1307,7 +1308,7 @@ using DeviceConvFwdInstances = std::tuple<
6, 1, // MRepeat x NRepeat
S<1, 96, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 16, 16, true, // ABlockTransferThreadClusterLengths_AK0_M_AK1, ArrangeOrder, SrcAccessOrder, VectorDim, SrcScalarPerVector, DstScalarPerVector_AK1, LdsExtraM
S<1, 48, 2>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, true,
1, 1, S<1, 32, 1, 3>, 16>,
6, 1, S<1, 32, 1, 3>, 16>,
ck::tensor_operation::device::DeviceGroupedConvFwdMultipleD_Wmma_CShuffle<
NDimSpatial,
InputLayout<NDimSpatial>, WeightLayout<NDimSpatial>, ck::Tuple<BiasLayout<NDimSpatial>, ResidualLayout<NDimSpatial>>, OutputLayout<NDimSpatial>,
......@@ -1318,7 +1319,7 @@ using DeviceConvFwdInstances = std::tuple<
6, 1, // MRepeat x NRepeat
S<1, 192, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 16, 16, true, // ABlockTransferThreadClusterLengths_AK0_M_AK1, ArrangeOrder, SrcAccessOrder, VectorDim, SrcScalarPerVector, DstScalarPerVector_AK1, LdsExtraM
S<1, 48, 4>, S<1, 0, 2>, S<1, 0, 2>, 2, 4, 4, true,
1, 1, S<1, 64, 1, 3>, 16>,
6, 1, S<1, 64, 1, 3>, 16>,
ck::tensor_operation::device::DeviceGroupedConvFwdMultipleD_Wmma_CShuffle<
NDimSpatial,
InputLayout<NDimSpatial>, WeightLayout<NDimSpatial>, ck::Tuple<BiasLayout<NDimSpatial>, ResidualLayout<NDimSpatial>>, OutputLayout<NDimSpatial>,
......@@ -1329,7 +1330,7 @@ using DeviceConvFwdInstances = std::tuple<
6, 1, // MRepeat x NRepeat
S<1, 96, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 16, 16, true, // ABlockTransferThreadClusterLengths_AK0_M_AK1, ArrangeOrder, SrcAccessOrder, VectorDim, SrcScalarPerVector, DstScalarPerVector_AK1, LdsExtraM
S<1, 48, 2>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, true,
1, 1, S<1, 32, 1, 3>, 16>,
6, 1, S<1, 32, 1, 3>, 16>,
ck::tensor_operation::device::DeviceGroupedConvFwdMultipleD_Wmma_CShuffle<
NDimSpatial,
InputLayout<NDimSpatial>, WeightLayout<NDimSpatial>, ck::Tuple<BiasLayout<NDimSpatial>, ResidualLayout<NDimSpatial>>, OutputLayout<NDimSpatial>,
......@@ -1340,7 +1341,7 @@ using DeviceConvFwdInstances = std::tuple<
6, 1, // MRepeat x NRepeat
S<1, 192, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 16, 16, true, // ABlockTransferThreadClusterLengths_AK0_M_AK1, ArrangeOrder, SrcAccessOrder, VectorDim, SrcScalarPerVector, DstScalarPerVector_AK1, LdsExtraM
S<1, 48, 4>, S<1, 0, 2>, S<1, 0, 2>, 2, 4, 4, true,
1, 1, S<1, 64, 1, 3>, 16>,
6, 1, S<1, 64, 1, 3>, 16>,
ck::tensor_operation::device::DeviceGroupedConvFwdMultipleD_Wmma_CShuffle<
NDimSpatial,
InputLayout<NDimSpatial>, WeightLayout<NDimSpatial>, ck::Tuple<BiasLayout<NDimSpatial>, ResidualLayout<NDimSpatial>>, OutputLayout<NDimSpatial>,
......@@ -1351,7 +1352,7 @@ using DeviceConvFwdInstances = std::tuple<
6, 1, // MRepeat x NRepeat
S<1, 96, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 16, 16, true, // ABlockTransferThreadClusterLengths_AK0_M_AK1, ArrangeOrder, SrcAccessOrder, VectorDim, SrcScalarPerVector, DstScalarPerVector_AK1, LdsExtraM
S<1, 48, 2>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, true,
1, 1, S<1, 32, 1, 3>, 16>,
6, 1, S<1, 32, 1, 3>, 16>,
ck::tensor_operation::device::DeviceGroupedConvFwdMultipleD_Wmma_CShuffle<
NDimSpatial,
InputLayout<NDimSpatial>, WeightLayout<NDimSpatial>, ck::Tuple<BiasLayout<NDimSpatial>, ResidualLayout<NDimSpatial>>, OutputLayout<NDimSpatial>,
......@@ -1362,7 +1363,7 @@ using DeviceConvFwdInstances = std::tuple<
6, 1, // MRepeat x NRepeat
S<1, 192, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 16, 16, true, // ABlockTransferThreadClusterLengths_AK0_M_AK1, ArrangeOrder, SrcAccessOrder, VectorDim, SrcScalarPerVector, DstScalarPerVector_AK1, LdsExtraM
S<1, 48, 4>, S<1, 0, 2>, S<1, 0, 2>, 2, 4, 4, true,
1, 1, S<1, 64, 1, 3>, 16>,
6, 1, S<1, 64, 1, 3>, 16>,
ck::tensor_operation::device::DeviceGroupedConvFwdMultipleD_Wmma_CShuffle<
NDimSpatial,
InputLayout<NDimSpatial>, WeightLayout<NDimSpatial>, ck::Tuple<BiasLayout<NDimSpatial>, ResidualLayout<NDimSpatial>>, OutputLayout<NDimSpatial>,
......@@ -1373,7 +1374,7 @@ using DeviceConvFwdInstances = std::tuple<
6, 1, // MRepeat x NRepeat
S<1, 96, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 16, 16, true, // ABlockTransferThreadClusterLengths_AK0_M_AK1, ArrangeOrder, SrcAccessOrder, VectorDim, SrcScalarPerVector, DstScalarPerVector_AK1, LdsExtraM
S<1, 48, 2>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, true,
1, 1, S<1, 32, 1, 3>, 16>,
6, 1, S<1, 32, 1, 3>, 16>,
ck::tensor_operation::device::DeviceGroupedConvFwdMultipleD_Wmma_CShuffle<
NDimSpatial,
InputLayout<NDimSpatial>, WeightLayout<NDimSpatial>, ck::Tuple<BiasLayout<NDimSpatial>, ResidualLayout<NDimSpatial>>, OutputLayout<NDimSpatial>,
......@@ -1384,7 +1385,7 @@ using DeviceConvFwdInstances = std::tuple<
6, 1, // MRepeat x NRepeat
S<1, 192, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 16, 16, true, // ABlockTransferThreadClusterLengths_AK0_M_AK1, ArrangeOrder, SrcAccessOrder, VectorDim, SrcScalarPerVector, DstScalarPerVector_AK1, LdsExtraM
S<1, 48, 4>, S<1, 0, 2>, S<1, 0, 2>, 2, 4, 4, true,
1, 1, S<1, 64, 1, 3>, 16>,
6, 1, S<1, 64, 1, 3>, 16>,
ck::tensor_operation::device::DeviceGroupedConvFwdMultipleD_Wmma_CShuffle<
NDimSpatial,
InputLayout<NDimSpatial>, WeightLayout<NDimSpatial>, ck::Tuple<BiasLayout<NDimSpatial>, ResidualLayout<NDimSpatial>>, OutputLayout<NDimSpatial>,
......@@ -1395,7 +1396,7 @@ using DeviceConvFwdInstances = std::tuple<
6, 1, // MRepeat x NRepeat
S<1, 96, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 16, 16, true, // ABlockTransferThreadClusterLengths_AK0_M_AK1, ArrangeOrder, SrcAccessOrder, VectorDim, SrcScalarPerVector, DstScalarPerVector_AK1, LdsExtraM
S<1, 48, 2>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, true,
1, 1, S<1, 32, 1, 3>, 16>,
6, 1, S<1, 32, 1, 3>, 16>,
ck::tensor_operation::device::DeviceGroupedConvFwdMultipleD_Wmma_CShuffle<
NDimSpatial,
InputLayout<NDimSpatial>, WeightLayout<NDimSpatial>, ck::Tuple<BiasLayout<NDimSpatial>, ResidualLayout<NDimSpatial>>, OutputLayout<NDimSpatial>,
......@@ -1406,7 +1407,7 @@ using DeviceConvFwdInstances = std::tuple<
6, 1, // MRepeat x NRepeat
S<1, 192, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 16, 16, true, // ABlockTransferThreadClusterLengths_AK0_M_AK1, ArrangeOrder, SrcAccessOrder, VectorDim, SrcScalarPerVector, DstScalarPerVector_AK1, LdsExtraM
S<1, 48, 4>, S<1, 0, 2>, S<1, 0, 2>, 2, 4, 4, true,
1, 1, S<1, 64, 1, 3>, 16>,
6, 1, S<1, 64, 1, 3>, 16>,
ck::tensor_operation::device::DeviceGroupedConvFwdMultipleD_Wmma_CShuffle<
NDimSpatial,
InputLayout<NDimSpatial>, WeightLayout<NDimSpatial>, ck::Tuple<BiasLayout<NDimSpatial>, ResidualLayout<NDimSpatial>>, OutputLayout<NDimSpatial>,
......@@ -1417,7 +1418,7 @@ using DeviceConvFwdInstances = std::tuple<
6, 1, // MRepeat x NRepeat
S<1, 96, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 16, 16, true, // ABlockTransferThreadClusterLengths_AK0_M_AK1, ArrangeOrder, SrcAccessOrder, VectorDim, SrcScalarPerVector, DstScalarPerVector_AK1, LdsExtraM
S<1, 48, 2>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, true,
1, 1, S<1, 32, 1, 3>, 16>,
6, 1, S<1, 32, 1, 3>, 16>,
ck::tensor_operation::device::DeviceGroupedConvFwdMultipleD_Wmma_CShuffle<
NDimSpatial,
InputLayout<NDimSpatial>, WeightLayout<NDimSpatial>, ck::Tuple<BiasLayout<NDimSpatial>, ResidualLayout<NDimSpatial>>, OutputLayout<NDimSpatial>,
......@@ -1428,7 +1429,7 @@ using DeviceConvFwdInstances = std::tuple<
6, 1, // MRepeat x NRepeat
S<1, 192, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 16, 16, true, // ABlockTransferThreadClusterLengths_AK0_M_AK1, ArrangeOrder, SrcAccessOrder, VectorDim, SrcScalarPerVector, DstScalarPerVector_AK1, LdsExtraM
S<1, 48, 4>, S<1, 0, 2>, S<1, 0, 2>, 2, 4, 4, true,
1, 1, S<1, 64, 1, 3>, 16>,
6, 1, S<1, 64, 1, 3>, 16>,
ck::tensor_operation::device::DeviceGroupedConvFwdMultipleD_Wmma_CShuffle<
NDimSpatial,
InputLayout<NDimSpatial>, WeightLayout<NDimSpatial>, ck::Tuple<BiasLayout<NDimSpatial>, ResidualLayout<NDimSpatial>>, OutputLayout<NDimSpatial>,
......@@ -1439,7 +1440,7 @@ using DeviceConvFwdInstances = std::tuple<
6, 1, // MRepeat x NRepeat
S<1, 96, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 16, 16, true, // ABlockTransferThreadClusterLengths_AK0_M_AK1, ArrangeOrder, SrcAccessOrder, VectorDim, SrcScalarPerVector, DstScalarPerVector_AK1, LdsExtraM
S<1, 48, 2>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, true,
1, 1, S<1, 32, 1, 3>, 16>
6, 1, S<1, 32, 1, 3>, 16>
>;
// 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