Merge remote-tracking branch 'origin/develop' into cpu_avx2

library/src/tensor_operation_instance/gpu/convnd_bwd_data/device_conv1d_bwd_data_xdl_nwc_kxc_nwk_int8_instance.cpp

+#include <stdlib.h>
+#include "config.hpp"
+#include "device_convnd_bwd_data_xdl_ndhwc_kzyxc_ndhwk.hpp"
+#include "element_wise_operation.hpp"
+#include "device_operation_instance.hpp"
+
+namespace ck {
+namespace tensor_operation {
+namespace device {
+namespace device_conv2d_bwd_data_instance {
+
+using DataType = int8_t;
+using AccType  = int32_t;
+
+template <ck::index_t... Is>
+using S = ck::Sequence<Is...>;
+
+using PassThrough = ck::tensor_operation::element_wise::PassThrough;
+static constexpr auto ConvBwdDataDefault =
+    ck::tensor_operation::device::ConvolutionBackwardDataSpecialization_t::Default;
+
+static constexpr auto ConvBwdDataFilter1x1Stride1Pad0 =
+    ck::tensor_operation::device::ConvolutionBackwardDataSpecialization_t::Filter1x1Stride1Pad0;
+
+// Compilation parameters for in[n, hi, wi, c] * wei[k, y, x, c] = out[n, ho, wo, k]
+using device_conv1d_bwd_data_xdl_nwc_kxc_nwk_int8_instances =
+    std::tuple<
+        // clang-format off
+        //#############################################################################|    InData|   WeiData|   OutData|    AccData|          In|         Wei|         Out|        ConvBackward|    Num| Block|  MPer|  NPer| K0Per|  K1| MPer| NPer| MXdl| NXdl|  ABlockTransfer| ABlockTransfer| ABlockTransfer| ABlockTransfer| ABlockTransfer| ABlockTransfer| ABlockLds|  BBlockTransfer| BBlockTransfer| BBlockTransfer| BlockTransfer| BBlockTransfer| BBlockTransfer| BBlockLds| CThreadTransfer| CThreadTransfer|
+        //#############################################################################|      Type|      Type|      Type|       Type| Elementwise| Elementwise| Elementwise|                Data|    Dim|  Size| Block| Block| Block|    |  XDL|  XDL|  Per|  Per|   ThreadCluster|  ThreadCluster| SrcAccessOrder|   SrcVectorDim|      SrcScalar|      DstScalar| AddExtraM|   ThreadCluster|  ThreadCluster| SrcAccessOrder|  SrcVectorDim|      SrcScalar|      DstScalar| AddExtraN| SrcDstVectorDim|       DstScalar|
+        //#############################################################################|          |          |          |           |   Operation|   Operation|   Operation|      Specialization|Spatial|      |      |      |      |    |     |     | Wave| Wave| Lengths_K0_M_K1|   ArrangeOrder|               |               |      PerVector|   PerVector_K1|          | Lengths_K0_N_K1|   ArrangeOrder|               |              |      PerVector|   PerVector_K1|          |                |       PerVector|
+        //#############################################################################|          |          |          |           |            |            |            |                    |       |      |      |      |      |    |     |     |     |     |                |               |               |               |               |               |          |                |               |               |              |               |               |          |                |                |
+        DeviceConvndBwdDataXdl_Input_N_Di_Hi_Wi_C_Weight_K_Z_Y_X_C_Output_N_Do_Ho_Wo_K<  DataType,  DataType,  DataType,    AccType, PassThrough, PassThrough, PassThrough,  ConvBwdDataDefault,   1,     256,   256,   128,     4,  16,   32,   32,    4,    2,     S<4, 64, 1>,     S<1, 0, 2>,     S<1, 0, 2>,              2,             16,             16,      true,     S<4, 64, 1>,     S<2, 0, 1>,     S<0, 2, 1>,             1,              2,             16,      true,               7,               1>,
+    #if 1
+        DeviceConvndBwdDataXdl_Input_N_Di_Hi_Wi_C_Weight_K_Z_Y_X_C_Output_N_Do_Ho_Wo_K<  DataType,  DataType,  DataType,    AccType, PassThrough, PassThrough, PassThrough,  ConvBwdDataDefault,   1,     256,   128,   256,     4,  16,   32,   32,    2,    4,     S<4, 64, 1>,     S<1, 0, 2>,     S<1, 0, 2>,              2,             16,             16,      true,     S<4, 64, 1>,     S<2, 0, 1>,     S<0, 2, 1>,             1,              4,             16,      true,               7,               1>,
+    #endif
+        DeviceConvndBwdDataXdl_Input_N_Di_Hi_Wi_C_Weight_K_Z_Y_X_C_Output_N_Do_Ho_Wo_K<  DataType,  DataType,  DataType,    AccType, PassThrough, PassThrough, PassThrough,  ConvBwdDataDefault,   1,     128,   128,   128,     4,  16,   32,   32,    4,    2,     S<4, 32, 1>,     S<1, 0, 2>,     S<1, 0, 2>,              2,             16,             16,      true,     S<4, 32, 1>,     S<2, 0, 1>,     S<0, 2, 1>,             1,              2,             16,      true,               7,               1>,
+        DeviceConvndBwdDataXdl_Input_N_Di_Hi_Wi_C_Weight_K_Z_Y_X_C_Output_N_Do_Ho_Wo_K<  DataType,  DataType,  DataType,    AccType, PassThrough, PassThrough, PassThrough,  ConvBwdDataDefault,   1,     256,   128,   128,     4,  16,   32,   32,    2,    2,     S<4, 64, 1>,     S<1, 0, 2>,     S<1, 0, 2>,              2,             16,             16,      true,     S<4, 64, 1>,     S<2, 0, 1>,     S<0, 2, 1>,             1,              2,             16,      true,               7,               1>,
+        DeviceConvndBwdDataXdl_Input_N_Di_Hi_Wi_C_Weight_K_Z_Y_X_C_Output_N_Do_Ho_Wo_K<  DataType,  DataType,  DataType,    AccType, PassThrough, PassThrough, PassThrough,  ConvBwdDataDefault,   1,     128,   128,    64,     4,  16,   32,   32,    2,    2,     S<4, 32, 1>,     S<1, 0, 2>,     S<1, 0, 2>,              2,             16,             16,      true,     S<4, 32, 1>,     S<2, 0, 1>,     S<0, 2, 1>,             1,              2,             16,      true,               7,               1>,
+        DeviceConvndBwdDataXdl_Input_N_Di_Hi_Wi_C_Weight_K_Z_Y_X_C_Output_N_Do_Ho_Wo_K<  DataType,  DataType,  DataType,    AccType, PassThrough, PassThrough, PassThrough,  ConvBwdDataDefault,   1,     128,    64,   128,     4,  16,   32,   32,    2,    2,     S<4, 32, 1>,     S<1, 0, 2>,     S<1, 0, 2>,              2,             16,             16,      true,     S<4, 32, 1>,     S<2, 0, 1>,     S<0, 2, 1>,             1,              4,             16,      true,               7,               1>,
+        DeviceConvndBwdDataXdl_Input_N_Di_Hi_Wi_C_Weight_K_Z_Y_X_C_Output_N_Do_Ho_Wo_K<  DataType,  DataType,  DataType,    AccType, PassThrough, PassThrough, PassThrough,  ConvBwdDataDefault,   1,      64,    64,    64,     4,  16,   32,   32,    2,    2,     S<4, 16, 1>,     S<1, 0, 2>,     S<1, 0, 2>,              2,             16,             16,      true,     S<4, 16, 1>,     S<2, 0, 1>,     S<0, 2, 1>,             1,              4,             16,      true,               7,               1>,
+        DeviceConvndBwdDataXdl_Input_N_Di_Hi_Wi_C_Weight_K_Z_Y_X_C_Output_N_Do_Ho_Wo_K<  DataType,  DataType,  DataType,    AccType, PassThrough, PassThrough, PassThrough,  ConvBwdDataDefault,   1,     256,   128,    64,     4,  16,   32,   32,    2,    1,     S<4, 64, 1>,     S<1, 0, 2>,     S<1, 0, 2>,              2,             16,             16,      true,     S<4, 64, 1>,     S<2, 0, 1>,     S<0, 2, 1>,             1,              1,             16,      true,               7,               1>,
+        DeviceConvndBwdDataXdl_Input_N_Di_Hi_Wi_C_Weight_K_Z_Y_X_C_Output_N_Do_Ho_Wo_K<  DataType,  DataType,  DataType,    AccType, PassThrough, PassThrough, PassThrough,  ConvBwdDataDefault,   1,     256,    64,   128,     4,  16,   32,   32,    1,    2,     S<4, 64, 1>,     S<1, 0, 2>,     S<1, 0, 2>,              2,             16,             16,      true,     S<4, 64, 1>,     S<2, 0, 1>,     S<0, 2, 1>,             1,              2,             16,      true,               7,               1>,
+        DeviceConvndBwdDataXdl_Input_N_Di_Hi_Wi_C_Weight_K_Z_Y_X_C_Output_N_Do_Ho_Wo_K<  DataType,  DataType,  DataType,    AccType, PassThrough, PassThrough, PassThrough,  ConvBwdDataDefault,   1,     128,   128,    32,     4,  16,   32,   32,    2,    1,     S<4, 32, 1>,     S<1, 0, 2>,     S<1, 0, 2>,              2,             16,             16,      true,     S<4, 32, 1>,     S<2, 0, 1>,     S<0, 2, 1>,             1,              1,             16,      true,               7,               1>,
+        DeviceConvndBwdDataXdl_Input_N_Di_Hi_Wi_C_Weight_K_Z_Y_X_C_Output_N_Do_Ho_Wo_K<  DataType,  DataType,  DataType,    AccType, PassThrough, PassThrough, PassThrough,  ConvBwdDataDefault,   1,     128,    32,   128,     4,  16,   32,   32,    1,    2,     S<4, 32, 1>,     S<1, 0, 2>,     S<1, 0, 2>,              2,             16,             16,      true,     S<4, 32, 1>,     S<2, 0, 1>,     S<0, 2, 1>,             1,              4,             16,      true,               7,               1>,
+        DeviceConvndBwdDataXdl_Input_N_Di_Hi_Wi_C_Weight_K_Z_Y_X_C_Output_N_Do_Ho_Wo_K<  DataType,  DataType,  DataType,    AccType, PassThrough, PassThrough, PassThrough,  ConvBwdDataDefault,   1,      64,    64,    32,     4,  16,   32,   32,    2,    1,     S<4, 16, 1>,     S<1, 0, 2>,     S<1, 0, 2>,              2,             16,             16,      true,     S<4, 16, 1>,     S<2, 0, 1>,     S<0, 2, 1>,             1,              2,             16,      true,               7,               1>,
+        DeviceConvndBwdDataXdl_Input_N_Di_Hi_Wi_C_Weight_K_Z_Y_X_C_Output_N_Do_Ho_Wo_K<  DataType,  DataType,  DataType,    AccType, PassThrough, PassThrough, PassThrough,  ConvBwdDataDefault,   1,      64,    32,    64,     4,  16,   32,   32,    1,    2,     S<4, 16, 1>,     S<1, 0, 2>,     S<1, 0, 2>,              2,             16,             16,      true,     S<4, 16, 1>,     S<2, 0, 1>,     S<0, 2, 1>,             1,              4,             16,      true,               7,               1>
+        // clang-format on
+        >;
+
+using device_conv1d_bwd_data_xdl_nwc_kxc_nwk_1x1_s1_p0_int8_instances =
+    std::tuple<
+        // clang-format off
+        //##############################################################################|    InData|   WeiData|   OutData|    AccData|          In|         Wei|         Out|                     ConvBackward|    Num| Block|  MPer|  NPer| K0Per|  K1| MPer| NPer| MXdl| NXdl|   ABlockTransfer| ABlockTransfer| ABlockTransfer| ABlockTransfer| ABlockTransfer| ABlockTransfer| ABlockLds|  BBlockTransfer| BBlockTransfer| BBlockTransfer| BlockTransfer| BBlockTransfer| BBlockTransfer| BBlockLds| CThreadTransfer| CThreadTransfer|
+        //##############################################################################|      Type|      Type|      Type|       Type| Elementwise| Elementwise| Elementwise|                             Data|    Dim|  Size| Block| Block| Block|    |  XDL|  XDL|  Per|  Per|    ThreadCluster|  ThreadCluster| SrcAccessOrder|   SrcVectorDim|      SrcScalar|      DstScalar| AddExtraM|   ThreadCluster|  ThreadCluster| SrcAccessOrder|  SrcVectorDim|      SrcScalar|      DstScalar| AddExtraN| SrcDstVectorDim|       DstScalar|
+        //##############################################################################|          |          |          |           |   Operation|   Operation|   Operation|                   Specialization|Spatial|      |      |      |      |    |     |     | Wave| Wave|  Lengths_K0_M_K1|   ArrangeOrder|               |               |      PerVector|   PerVector_K1|          | Lengths_K0_N_K1|   ArrangeOrder|               |              |      PerVector|   PerVector_K1|          |                |       PerVector|
+        //##############################################################################|          |          |          |           |            |            |            |                                 |       |      |      |      |      |    |     |     |     |     |                 |               |               |               |               |               |          |                |               |               |              |               |               |          |                |                |
+        DeviceConvndBwdDataXdl_Input_N_Di_Hi_Wi_C_Weight_K_Z_Y_X_C_Output_N_Do_Ho_Wo_K<   DataType,  DataType,  DataType,    AccType, PassThrough, PassThrough, PassThrough,  ConvBwdDataFilter1x1Stride1Pad0,   1,     256,   256,   128,     4,  16,   32,   32,    4,    2,      S<4, 64, 1>,     S<1, 0, 2>,     S<1, 0, 2>,              2,             16,             16,      true,     S<4, 64, 1>,     S<2, 0, 1>,     S<0, 2, 1>,             1,              2,             16,      true,               7,               1>,
+        DeviceConvndBwdDataXdl_Input_N_Di_Hi_Wi_C_Weight_K_Z_Y_X_C_Output_N_Do_Ho_Wo_K<   DataType,  DataType,  DataType,    AccType, PassThrough, PassThrough, PassThrough,  ConvBwdDataFilter1x1Stride1Pad0,   1,     256,   128,   256,     4,  16,   32,   32,    2,    4,      S<4, 64, 1>,     S<1, 0, 2>,     S<1, 0, 2>,              2,             16,             16,      true,     S<4, 64, 1>,     S<2, 0, 1>,     S<0, 2, 1>,             1,              4,             16,      true,               7,               1>,
+        DeviceConvndBwdDataXdl_Input_N_Di_Hi_Wi_C_Weight_K_Z_Y_X_C_Output_N_Do_Ho_Wo_K<   DataType,  DataType,  DataType,    AccType, PassThrough, PassThrough, PassThrough,  ConvBwdDataFilter1x1Stride1Pad0,   1,     128,   128,   128,     4,  16,   32,   32,    4,    2,      S<4, 32, 1>,     S<1, 0, 2>,     S<1, 0, 2>,              2,             16,             16,      true,     S<4, 32, 1>,     S<2, 0, 1>,     S<0, 2, 1>,             1,              2,             16,      true,               7,               1>,
+        DeviceConvndBwdDataXdl_Input_N_Di_Hi_Wi_C_Weight_K_Z_Y_X_C_Output_N_Do_Ho_Wo_K<   DataType,  DataType,  DataType,    AccType, PassThrough, PassThrough, PassThrough,  ConvBwdDataFilter1x1Stride1Pad0,   1,     256,   128,   128,     4,  16,   32,   32,    2,    2,      S<4, 64, 1>,     S<1, 0, 2>,     S<1, 0, 2>,              2,             16,             16,      true,     S<4, 64, 1>,     S<2, 0, 1>,     S<0, 2, 1>,             1,              2,             16,      true,               7,               1>,
+        DeviceConvndBwdDataXdl_Input_N_Di_Hi_Wi_C_Weight_K_Z_Y_X_C_Output_N_Do_Ho_Wo_K<   DataType,  DataType,  DataType,    AccType, PassThrough, PassThrough, PassThrough,  ConvBwdDataFilter1x1Stride1Pad0,   1,     128,   128,    64,     4,  16,   32,   32,    2,    2,      S<4, 32, 1>,     S<1, 0, 2>,     S<1, 0, 2>,              2,             16,             16,      true,     S<4, 32, 1>,     S<2, 0, 1>,     S<0, 2, 1>,             1,              2,             16,      true,               7,               1>,
+        DeviceConvndBwdDataXdl_Input_N_Di_Hi_Wi_C_Weight_K_Z_Y_X_C_Output_N_Do_Ho_Wo_K<   DataType,  DataType,  DataType,    AccType, PassThrough, PassThrough, PassThrough,  ConvBwdDataFilter1x1Stride1Pad0,   1,     128,    64,   128,     4,  16,   32,   32,    2,    2,      S<4, 32, 1>,     S<1, 0, 2>,     S<1, 0, 2>,              2,             16,             16,      true,     S<4, 32, 1>,     S<2, 0, 1>,     S<0, 2, 1>,             1,              4,             16,      true,               7,               1>,
+        DeviceConvndBwdDataXdl_Input_N_Di_Hi_Wi_C_Weight_K_Z_Y_X_C_Output_N_Do_Ho_Wo_K<   DataType,  DataType,  DataType,    AccType, PassThrough, PassThrough, PassThrough,  ConvBwdDataFilter1x1Stride1Pad0,   1,      64,    64,    64,     4,  16,   32,   32,    2,    2,      S<4, 16, 1>,     S<1, 0, 2>,     S<1, 0, 2>,              2,             16,             16,      true,     S<4, 16, 1>,     S<2, 0, 1>,     S<0, 2, 1>,             1,              4,             16,      true,               7,               1>,
+        DeviceConvndBwdDataXdl_Input_N_Di_Hi_Wi_C_Weight_K_Z_Y_X_C_Output_N_Do_Ho_Wo_K<   DataType,  DataType,  DataType,    AccType, PassThrough, PassThrough, PassThrough,  ConvBwdDataFilter1x1Stride1Pad0,   1,     256,   128,    64,     4,  16,   32,   32,    2,    1,      S<4, 64, 1>,     S<1, 0, 2>,     S<1, 0, 2>,              2,             16,             16,      true,     S<4, 64, 1>,     S<2, 0, 1>,     S<0, 2, 1>,             1,              1,             16,      true,               7,               1>,
+        DeviceConvndBwdDataXdl_Input_N_Di_Hi_Wi_C_Weight_K_Z_Y_X_C_Output_N_Do_Ho_Wo_K<   DataType,  DataType,  DataType,    AccType, PassThrough, PassThrough, PassThrough,  ConvBwdDataFilter1x1Stride1Pad0,   1,     256,    64,   128,     4,  16,   32,   32,    1,    2,      S<4, 64, 1>,     S<1, 0, 2>,     S<1, 0, 2>,              2,             16,             16,      true,     S<4, 64, 1>,     S<2, 0, 1>,     S<0, 2, 1>,             1,              2,             16,      true,               7,               1>,
+        DeviceConvndBwdDataXdl_Input_N_Di_Hi_Wi_C_Weight_K_Z_Y_X_C_Output_N_Do_Ho_Wo_K<   DataType,  DataType,  DataType,    AccType, PassThrough, PassThrough, PassThrough,  ConvBwdDataFilter1x1Stride1Pad0,   1,     128,   128,    32,     4,  16,   32,   32,    2,    1,      S<4, 32, 1>,     S<1, 0, 2>,     S<1, 0, 2>,              2,             16,             16,      true,     S<4, 32, 1>,     S<2, 0, 1>,     S<0, 2, 1>,             1,              1,             16,      true,               7,               1>,
+        DeviceConvndBwdDataXdl_Input_N_Di_Hi_Wi_C_Weight_K_Z_Y_X_C_Output_N_Do_Ho_Wo_K<   DataType,  DataType,  DataType,    AccType, PassThrough, PassThrough, PassThrough,  ConvBwdDataFilter1x1Stride1Pad0,   1,     128,    32,   128,     4,  16,   32,   32,    1,    2,      S<4, 32, 1>,     S<1, 0, 2>,     S<1, 0, 2>,              2,             16,             16,      true,     S<4, 32, 1>,     S<2, 0, 1>,     S<0, 2, 1>,             1,              4,             16,      true,               7,               1>,
+        DeviceConvndBwdDataXdl_Input_N_Di_Hi_Wi_C_Weight_K_Z_Y_X_C_Output_N_Do_Ho_Wo_K<   DataType,  DataType,  DataType,    AccType, PassThrough, PassThrough, PassThrough,  ConvBwdDataFilter1x1Stride1Pad0,   1,      64,    64,    32,     4,  16,   32,   32,    2,    1,      S<4, 16, 1>,     S<1, 0, 2>,     S<1, 0, 2>,              2,             16,             16,      true,     S<4, 16, 1>,     S<2, 0, 1>,     S<0, 2, 1>,             1,              2,             16,      true,               7,               1>,
+        DeviceConvndBwdDataXdl_Input_N_Di_Hi_Wi_C_Weight_K_Z_Y_X_C_Output_N_Do_Ho_Wo_K<   DataType,  DataType,  DataType,    AccType, PassThrough, PassThrough, PassThrough,  ConvBwdDataFilter1x1Stride1Pad0,   1,      64,    32,    64,     4,  16,   32,   32,    1,    2,      S<4, 16, 1>,     S<1, 0, 2>,     S<1, 0, 2>,              2,             16,             16,      true,     S<4, 16, 1>,     S<2, 0, 1>,     S<0, 2, 1>,             1,              4,             16,      true,               7,               1>
+        // clang-format on
+        >;
+
+void add_device_conv1d_bwd_data_xdl_nwc_kxc_nwk_int8_instances(
+    std::vector<DeviceConvBwdDataPtr<PassThrough, PassThrough, PassThrough>>& instances)
+{
+    add_device_operation_instances(instances,
+                                   device_conv1d_bwd_data_xdl_nwc_kxc_nwk_int8_instances{});
+    add_device_operation_instances(
+        instances, device_conv1d_bwd_data_xdl_nwc_kxc_nwk_1x1_s1_p0_int8_instances{});
+}
+
+} // namespace device_conv2d_bwd_data_instance
+} // namespace device
+} // namespace tensor_operation
+} // namespace ck

library/src/tensor_operation_instance/gpu/convnd_bwd_data/device_conv2d_bwd_data_xdl_nhwc_kyxc_nhwk_bf16_instance.cpp

+#include <stdlib.h>
+#include "config.hpp"
+#include "device_convnd_bwd_data_xdl_ndhwc_kzyxc_ndhwk.hpp"
+#include "element_wise_operation.hpp"
+#include "device_operation_instance.hpp"
+
+namespace ck {
+namespace tensor_operation {
+namespace device {
+namespace device_conv2d_bwd_data_instance {
+
+using BF16 = ushort;
+using F32  = float;
+
+template <ck::index_t... Is>
+using S = ck::Sequence<Is...>;
+
+using PassThrough = ck::tensor_operation::element_wise::PassThrough;
+static constexpr auto ConvBwdDataDefault =
+    ck::tensor_operation::device::ConvolutionBackwardDataSpecialization_t::Default;
+
+static constexpr auto ConvBwdDataFilter1x1Stride1Pad0 =
+    ck::tensor_operation::device::ConvolutionBackwardDataSpecialization_t::Filter1x1Stride1Pad0;
+
+// Compilation parameters for in[n, hi, wi, c] * wei[k, y, x, c] = out[n, ho, wo, k]
+using device_conv2d_bwd_data_xdl_nhwc_kyxc_nhwk_bf16_instances =
+    std::tuple<
+        // clang-format off
+        //#############################################################################| InData| WeiData| OutData| AccData|          In|         Wei|         Out|        ConvBackward|    Num| Block|  MPer|  NPer| K0Per| K1| MPer| NPer| MXdl| NXdl|  ABlockTransfer| ABlockTransfer| ABlockTransfer| ABlockTransfer| ABlockTransfer| ABlockTransfer| ABlockLds|  BBlockTransfer| BBlockTransfer| BBlockTransfer| BlockTransfer| BBlockTransfer| BBlockTransfer| BBlockLds| CThreadTransfer| CThreadTransfer|
+        //#############################################################################|   Type|    Type|    Type|    Type| Elementwise| Elementwise| Elementwise|                Data|    Dim|  Size| Block| Block| Block|   |  XDL|  XDL|  Per|  Per|   ThreadCluster|  ThreadCluster| SrcAccessOrder|   SrcVectorDim|      SrcScalar|      DstScalar| AddExtraM|   ThreadCluster|  ThreadCluster| SrcAccessOrder|  SrcVectorDim|      SrcScalar|      DstScalar| AddExtraN| SrcDstVectorDim|       DstScalar|
+        //#############################################################################|       |        |        |        |   Operation|   Operation|   Operation|      Specialization|Spatial|      |      |      |      |   |     |     | Wave| Wave| Lengths_K0_M_K1|   ArrangeOrder|               |               |      PerVector|   PerVector_K1|          | Lengths_K0_N_K1|   ArrangeOrder|               |              |      PerVector|   PerVector_K1|          |                |       PerVector|
+        //#############################################################################|       |        |        |        |            |            |            |                    |       |      |      |      |      |   |     |     |     |     |                |               |               |               |               |               |          |                |               |               |              |               |               |          |                |                |
+        DeviceConvndBwdDataXdl_Input_N_Di_Hi_Wi_C_Weight_K_Z_Y_X_C_Output_N_Do_Ho_Wo_K<   BF16,    BF16,    BF16,     F32, PassThrough, PassThrough, PassThrough,  ConvBwdDataDefault,   2,     256,   256,   128,     4,  8,   32,   32,    4,    2,     S<4, 64, 1>,     S<1, 0, 2>,     S<1, 0, 2>,              2,              8,              8,      true,     S<4, 64, 1>,     S<2, 0, 1>,     S<0, 2, 1>,             1,              2,              8,      true,               7,               1>,
+        DeviceConvndBwdDataXdl_Input_N_Di_Hi_Wi_C_Weight_K_Z_Y_X_C_Output_N_Do_Ho_Wo_K<   BF16,    BF16,    BF16,     F32, PassThrough, PassThrough, PassThrough,  ConvBwdDataDefault,   2,     256,   128,   256,     4,  8,   32,   32,    2,    4,     S<4, 64, 1>,     S<1, 0, 2>,     S<1, 0, 2>,              2,              8,              8,      true,     S<4, 64, 1>,     S<2, 0, 1>,     S<0, 2, 1>,             1,              4,              8,      true,               7,               1>,
+        DeviceConvndBwdDataXdl_Input_N_Di_Hi_Wi_C_Weight_K_Z_Y_X_C_Output_N_Do_Ho_Wo_K<   BF16,    BF16,    BF16,     F32, PassThrough, PassThrough, PassThrough,  ConvBwdDataDefault,   2,     128,   128,   128,     4,  8,   32,   32,    4,    2,     S<4, 32, 1>,     S<1, 0, 2>,     S<1, 0, 2>,              2,              8,              8,      true,     S<4, 32, 1>,     S<2, 0, 1>,     S<0, 2, 1>,             1,              4,              8,      true,               7,               1>,
+        DeviceConvndBwdDataXdl_Input_N_Di_Hi_Wi_C_Weight_K_Z_Y_X_C_Output_N_Do_Ho_Wo_K<   BF16,    BF16,    BF16,     F32, PassThrough, PassThrough, PassThrough,  ConvBwdDataDefault,   2,     256,   128,   128,     4,  8,   32,   32,    2,    2,     S<4, 64, 1>,     S<1, 0, 2>,     S<1, 0, 2>,              2,              8,              8,      true,     S<4, 64, 1>,     S<2, 0, 1>,     S<0, 2, 1>,             1,              2,              8,      true,               7,               1>,
+        DeviceConvndBwdDataXdl_Input_N_Di_Hi_Wi_C_Weight_K_Z_Y_X_C_Output_N_Do_Ho_Wo_K<   BF16,    BF16,    BF16,     F32, PassThrough, PassThrough, PassThrough,  ConvBwdDataDefault,   2,     128,   128,    64,     4,  8,   32,   32,    2,    2,     S<4, 32, 1>,     S<1, 0, 2>,     S<1, 0, 2>,              2,              8,              8,      true,     S<4, 32, 1>,     S<2, 0, 1>,     S<0, 2, 1>,             1,              2,              8,      true,               7,               1>,
+        DeviceConvndBwdDataXdl_Input_N_Di_Hi_Wi_C_Weight_K_Z_Y_X_C_Output_N_Do_Ho_Wo_K<   BF16,    BF16,    BF16,     F32, PassThrough, PassThrough, PassThrough,  ConvBwdDataDefault,   2,     128,    64,   128,     4,  8,   32,   32,    2,    2,     S<4, 32, 1>,     S<1, 0, 2>,     S<1, 0, 2>,              2,              8,              8,      true,     S<4, 32, 1>,     S<2, 0, 1>,     S<0, 2, 1>,             1,              4,              8,      true,               7,               1>,
+        DeviceConvndBwdDataXdl_Input_N_Di_Hi_Wi_C_Weight_K_Z_Y_X_C_Output_N_Do_Ho_Wo_K<   BF16,    BF16,    BF16,     F32, PassThrough, PassThrough, PassThrough,  ConvBwdDataDefault,   2,      64,    64,    64,     4,  8,   32,   32,    2,    2,     S<4, 16, 1>,     S<1, 0, 2>,     S<1, 0, 2>,              2,              8,              8,      true,     S<4, 16, 1>,     S<2, 0, 1>,     S<0, 2, 1>,             1,              4,              8,      true,               7,               1>,
+        DeviceConvndBwdDataXdl_Input_N_Di_Hi_Wi_C_Weight_K_Z_Y_X_C_Output_N_Do_Ho_Wo_K<   BF16,    BF16,    BF16,     F32, PassThrough, PassThrough, PassThrough,  ConvBwdDataDefault,   2,     256,   128,    64,     4,  8,   32,   32,    2,    1,     S<4, 64, 1>,     S<1, 0, 2>,     S<1, 0, 2>,              2,              8,              8,      true,     S<4, 64, 1>,     S<2, 0, 1>,     S<0, 2, 1>,             1,              1,              8,      true,               7,               1>,
+        DeviceConvndBwdDataXdl_Input_N_Di_Hi_Wi_C_Weight_K_Z_Y_X_C_Output_N_Do_Ho_Wo_K<   BF16,    BF16,    BF16,     F32, PassThrough, PassThrough, PassThrough,  ConvBwdDataDefault,   2,     256,    64,   128,     4,  8,   32,   32,    1,    2,     S<4, 64, 1>,     S<1, 0, 2>,     S<1, 0, 2>,              2,              8,              8,      true,     S<4, 64, 1>,     S<2, 0, 1>,     S<0, 2, 1>,             1,              2,              8,      true,               7,               1>,
+        DeviceConvndBwdDataXdl_Input_N_Di_Hi_Wi_C_Weight_K_Z_Y_X_C_Output_N_Do_Ho_Wo_K<   BF16,    BF16,    BF16,     F32, PassThrough, PassThrough, PassThrough,  ConvBwdDataDefault,   2,     128,   128,    32,     4,  8,   32,   32,    2,    1,     S<4, 32, 1>,     S<1, 0, 2>,     S<1, 0, 2>,              2,              8,              8,      true,     S<4, 32, 1>,     S<2, 0, 1>,     S<0, 2, 1>,             1,              1,              8,      true,               7,               1>,
+        DeviceConvndBwdDataXdl_Input_N_Di_Hi_Wi_C_Weight_K_Z_Y_X_C_Output_N_Do_Ho_Wo_K<   BF16,    BF16,    BF16,     F32, PassThrough, PassThrough, PassThrough,  ConvBwdDataDefault,   2,     128,    32,   128,     4,  8,   32,   32,    1,    2,     S<4, 32, 1>,     S<1, 0, 2>,     S<1, 0, 2>,              2,              8,              8,      true,     S<4, 32, 1>,     S<2, 0, 1>,     S<0, 2, 1>,             1,              4,              8,      true,               7,               1>,
+        DeviceConvndBwdDataXdl_Input_N_Di_Hi_Wi_C_Weight_K_Z_Y_X_C_Output_N_Do_Ho_Wo_K<   BF16,    BF16,    BF16,     F32, PassThrough, PassThrough, PassThrough,  ConvBwdDataDefault,   2,      64,    64,    32,     4,  8,   32,   32,    2,    1,     S<4, 16, 1>,     S<1, 0, 2>,     S<1, 0, 2>,              2,              8,              8,      true,     S<4, 16, 1>,     S<2, 0, 1>,     S<0, 2, 1>,             1,              2,              8,      true,               7,               1>,
+        DeviceConvndBwdDataXdl_Input_N_Di_Hi_Wi_C_Weight_K_Z_Y_X_C_Output_N_Do_Ho_Wo_K<   BF16,    BF16,    BF16,     F32, PassThrough, PassThrough, PassThrough,  ConvBwdDataDefault,   2,      64,    32,    64,     4,  8,   32,   32,    1,    2,     S<4, 16, 1>,     S<1, 0, 2>,     S<1, 0, 2>,              2,              8,              8,      true,     S<4, 16, 1>,     S<2, 0, 1>,     S<0, 2, 1>,             1,              2,              8,      true,               7,               1>
+        // clang-format on
+        >;
+
+using device_conv2d_bwd_data_xdl_nhwc_kyxc_nhwk_1x1_s1_p0_bf16_instances =
+    std::tuple<
+        // clang-format off
+        //#############################################################################| InData| WeiData| OutData| AccData|          In|         Wei|         Out|                     ConvBackward|    Num| Block|  MPer|  NPer| K0Per| K1| MPer| NPer| MXdl| NXdl|  ABlockTransfer| ABlockTransfer| ABlockTransfer| ABlockTransfer| ABlockTransfer| ABlockTransfer| ABlockLds|  BBlockTransfer| BBlockTransfer| BBlockTransfer| BlockTransfer| BBlockTransfer| BBlockTransfer| BBlockLds| CThreadTransfer| CThreadTransfer|
+        //#############################################################################|   Type|    Type|    Type|    Type| Elementwise| Elementwise| Elementwise|                             Data|    Dim|  Size| Block| Block| Block|   |  XDL|  XDL|  Per|  Per|   ThreadCluster|  ThreadCluster| SrcAccessOrder|   SrcVectorDim|      SrcScalar|      DstScalar| AddExtraM|   ThreadCluster|  ThreadCluster| SrcAccessOrder|  SrcVectorDim|      SrcScalar|      DstScalar| AddExtraN| SrcDstVectorDim|       DstScalar|
+        //#############################################################################|       |        |        |        |   Operation|   Operation|   Operation|                   Specialization|Spatial|      |      |      |      |   |     |     | Wave| Wave| Lengths_K0_M_K1|   ArrangeOrder|               |               |      PerVector|   PerVector_K1|          | Lengths_K0_N_K1|   ArrangeOrder|               |              |      PerVector|   PerVector_K1|          |                |       PerVector|
+        //#############################################################################|       |        |        |        |            |            |            |                                 |       |      |      |      |      |   |     |     |     |     |                |               |               |               |               |               |          |                |               |               |              |               |               |          |                |                |
+        DeviceConvndBwdDataXdl_Input_N_Di_Hi_Wi_C_Weight_K_Z_Y_X_C_Output_N_Do_Ho_Wo_K<   BF16,    BF16,    BF16,     F32, PassThrough, PassThrough, PassThrough,  ConvBwdDataFilter1x1Stride1Pad0,   2,     256,   256,   128,     4,  8,   32,   32,    4,    2,     S<4, 64, 1>,     S<1, 0, 2>,     S<1, 0, 2>,              2,              8,              8,      true,     S<4, 64, 1>,     S<2, 0, 1>,     S<0, 2, 1>,             1,              2,              8,      true,               7,               1>,
+        DeviceConvndBwdDataXdl_Input_N_Di_Hi_Wi_C_Weight_K_Z_Y_X_C_Output_N_Do_Ho_Wo_K<   BF16,    BF16,    BF16,     F32, PassThrough, PassThrough, PassThrough,  ConvBwdDataFilter1x1Stride1Pad0,   2,     256,   128,   256,     4,  8,   32,   32,    2,    4,     S<4, 64, 1>,     S<1, 0, 2>,     S<1, 0, 2>,              2,              8,              8,      true,     S<4, 64, 1>,     S<2, 0, 1>,     S<0, 2, 1>,             1,              4,              8,      true,               7,               1>,
+        DeviceConvndBwdDataXdl_Input_N_Di_Hi_Wi_C_Weight_K_Z_Y_X_C_Output_N_Do_Ho_Wo_K<   BF16,    BF16,    BF16,     F32, PassThrough, PassThrough, PassThrough,  ConvBwdDataFilter1x1Stride1Pad0,   2,     128,   128,   128,     4,  8,   32,   32,    4,    2,     S<4, 32, 1>,     S<1, 0, 2>,     S<1, 0, 2>,              2,              8,              8,      true,     S<4, 32, 1>,     S<2, 0, 1>,     S<0, 2, 1>,             1,              4,              8,      true,               7,               1>,
+        DeviceConvndBwdDataXdl_Input_N_Di_Hi_Wi_C_Weight_K_Z_Y_X_C_Output_N_Do_Ho_Wo_K<   BF16,    BF16,    BF16,     F32, PassThrough, PassThrough, PassThrough,  ConvBwdDataFilter1x1Stride1Pad0,   2,     256,   128,   128,     4,  8,   32,   32,    2,    2,     S<4, 64, 1>,     S<1, 0, 2>,     S<1, 0, 2>,              2,              8,              8,      true,     S<4, 64, 1>,     S<2, 0, 1>,     S<0, 2, 1>,             1,              2,              8,      true,               7,               1>,
+        DeviceConvndBwdDataXdl_Input_N_Di_Hi_Wi_C_Weight_K_Z_Y_X_C_Output_N_Do_Ho_Wo_K<   BF16,    BF16,    BF16,     F32, PassThrough, PassThrough, PassThrough,  ConvBwdDataFilter1x1Stride1Pad0,   2,     128,   128,    64,     4,  8,   32,   32,    2,    2,     S<4, 32, 1>,     S<1, 0, 2>,     S<1, 0, 2>,              2,              8,              8,      true,     S<4, 32, 1>,     S<2, 0, 1>,     S<0, 2, 1>,             1,              2,              8,      true,               7,               1>,
+        DeviceConvndBwdDataXdl_Input_N_Di_Hi_Wi_C_Weight_K_Z_Y_X_C_Output_N_Do_Ho_Wo_K<   BF16,    BF16,    BF16,     F32, PassThrough, PassThrough, PassThrough,  ConvBwdDataFilter1x1Stride1Pad0,   2,     128,    64,   128,     4,  8,   32,   32,    2,    2,     S<4, 32, 1>,     S<1, 0, 2>,     S<1, 0, 2>,              2,              8,              8,      true,     S<4, 32, 1>,     S<2, 0, 1>,     S<0, 2, 1>,             1,              4,              8,      true,               7,               1>,
+        DeviceConvndBwdDataXdl_Input_N_Di_Hi_Wi_C_Weight_K_Z_Y_X_C_Output_N_Do_Ho_Wo_K<   BF16,    BF16,    BF16,     F32, PassThrough, PassThrough, PassThrough,  ConvBwdDataFilter1x1Stride1Pad0,   2,      64,    64,    64,     4,  8,   32,   32,    2,    2,     S<4, 16, 1>,     S<1, 0, 2>,     S<1, 0, 2>,              2,              8,              8,      true,     S<4, 16, 1>,     S<2, 0, 1>,     S<0, 2, 1>,             1,              4,              8,      true,               7,               1>,
+        DeviceConvndBwdDataXdl_Input_N_Di_Hi_Wi_C_Weight_K_Z_Y_X_C_Output_N_Do_Ho_Wo_K<   BF16,    BF16,    BF16,     F32, PassThrough, PassThrough, PassThrough,  ConvBwdDataFilter1x1Stride1Pad0,   2,     256,   128,    64,     4,  8,   32,   32,    2,    1,     S<4, 64, 1>,     S<1, 0, 2>,     S<1, 0, 2>,              2,              8,              8,      true,     S<4, 64, 1>,     S<2, 0, 1>,     S<0, 2, 1>,             1,              1,              8,      true,               7,               1>,
+        DeviceConvndBwdDataXdl_Input_N_Di_Hi_Wi_C_Weight_K_Z_Y_X_C_Output_N_Do_Ho_Wo_K<   BF16,    BF16,    BF16,     F32, PassThrough, PassThrough, PassThrough,  ConvBwdDataFilter1x1Stride1Pad0,   2,     256,    64,   128,     4,  8,   32,   32,    1,    2,     S<4, 64, 1>,     S<1, 0, 2>,     S<1, 0, 2>,              2,              8,              8,      true,     S<4, 64, 1>,     S<2, 0, 1>,     S<0, 2, 1>,             1,              2,              8,      true,               7,               1>,
+        DeviceConvndBwdDataXdl_Input_N_Di_Hi_Wi_C_Weight_K_Z_Y_X_C_Output_N_Do_Ho_Wo_K<   BF16,    BF16,    BF16,     F32, PassThrough, PassThrough, PassThrough,  ConvBwdDataFilter1x1Stride1Pad0,   2,     128,   128,    32,     4,  8,   32,   32,    2,    1,     S<4, 32, 1>,     S<1, 0, 2>,     S<1, 0, 2>,              2,              8,              8,      true,     S<4, 32, 1>,     S<2, 0, 1>,     S<0, 2, 1>,             1,              1,              8,      true,               7,               1>,
+        DeviceConvndBwdDataXdl_Input_N_Di_Hi_Wi_C_Weight_K_Z_Y_X_C_Output_N_Do_Ho_Wo_K<   BF16,    BF16,    BF16,     F32, PassThrough, PassThrough, PassThrough,  ConvBwdDataFilter1x1Stride1Pad0,   2,     128,    32,   128,     4,  8,   32,   32,    1,    2,     S<4, 32, 1>,     S<1, 0, 2>,     S<1, 0, 2>,              2,              8,              8,      true,     S<4, 32, 1>,     S<2, 0, 1>,     S<0, 2, 1>,             1,              4,              8,      true,               7,               1>,
+        DeviceConvndBwdDataXdl_Input_N_Di_Hi_Wi_C_Weight_K_Z_Y_X_C_Output_N_Do_Ho_Wo_K<   BF16,    BF16,    BF16,     F32, PassThrough, PassThrough, PassThrough,  ConvBwdDataFilter1x1Stride1Pad0,   2,      64,    64,    32,     4,  8,   32,   32,    2,    1,     S<4, 16, 1>,     S<1, 0, 2>,     S<1, 0, 2>,              2,              8,              8,      true,     S<4, 16, 1>,     S<2, 0, 1>,     S<0, 2, 1>,             1,              2,              8,      true,               7,               1>,
+        DeviceConvndBwdDataXdl_Input_N_Di_Hi_Wi_C_Weight_K_Z_Y_X_C_Output_N_Do_Ho_Wo_K<   BF16,    BF16,    BF16,     F32, PassThrough, PassThrough, PassThrough,  ConvBwdDataFilter1x1Stride1Pad0,   2,      64,    32,    64,     4,  8,   32,   32,    1,    2,     S<4, 16, 1>,     S<1, 0, 2>,     S<1, 0, 2>,              2,              8,              8,      true,     S<4, 16, 1>,     S<2, 0, 1>,     S<0, 2, 1>,             1,              2,              8,      true,               7,               1>
+        // clang-format on
+        >;
+
+void add_device_conv2d_bwd_data_xdl_nhwc_kyxc_nhwk_bf16_instances(
+    std::vector<DeviceConvBwdDataPtr<PassThrough, PassThrough, PassThrough>>& instances)
+{
+    add_device_operation_instances(instances,
+                                   device_conv2d_bwd_data_xdl_nhwc_kyxc_nhwk_bf16_instances{});
+    add_device_operation_instances(
+        instances, device_conv2d_bwd_data_xdl_nhwc_kyxc_nhwk_1x1_s1_p0_bf16_instances{});
+}
+
+} // namespace device_conv2d_bwd_data_instance
+} // namespace device
+} // namespace tensor_operation
+} // namespace ck

library/src/tensor_operation_instance/gpu/convnd_bwd_data/device_conv2d_bwd_data_xdl_nhwc_kyxc_nhwk_f16_instance.cpp

+#include <stdlib.h>
+#include "config.hpp"
+#include "device_convnd_bwd_data_xdl_ndhwc_kzyxc_ndhwk.hpp"
+#include "element_wise_operation.hpp"
+#include "device_operation_instance.hpp"
+
+namespace ck {
+namespace tensor_operation {
+namespace device {
+namespace device_conv2d_bwd_data_instance {
+
+using F16 = ck::half_t;
+using F32 = float;
+
+template <ck::index_t... Is>
+using S = ck::Sequence<Is...>;
+
+using PassThrough = ck::tensor_operation::element_wise::PassThrough;
+static constexpr auto ConvBwdDataDefault =
+    ck::tensor_operation::device::ConvolutionBackwardDataSpecialization_t::Default;
+
+static constexpr auto ConvBwdDataFilter1x1Stride1Pad0 =
+    ck::tensor_operation::device::ConvolutionBackwardDataSpecialization_t::Filter1x1Stride1Pad0;
+
+// Compilation parameters for in[n, hi, wi, c] * wei[k, y, x, c] = out[n, ho, wo, k]
+using device_conv2d_bwd_data_xdl_nhwc_kyxc_nhwk_f16_instances =
+    std::tuple<
+        // clang-format off
+        //#############################################################################| InData| WeiData| OutData| AccData|          In|         Wei|         Out|       ConvBackward|    Num| Block|  MPer|  NPer| K0Per| K1| MPer| NPer| MXdl| NXdl|  ABlockTransfer| ABlockTransfer| ABlockTransfer| ABlockTransfer| ABlockTransfer| ABlockTransfer| ABlockLds|  BBlockTransfer| BBlockTransfer| BBlockTransfer| BlockTransfer| BBlockTransfer| BBlockTransfer| BBlockLds| CThreadTransfer| CThreadTransfer|
+        //#############################################################################|   Type|    Type|    Type|    Type| Elementwise| Elementwise| Elementwise|               Data|    Dim|  Size| Block| Block| Block|   |  XDL|  XDL|  Per|  Per|   ThreadCluster|  ThreadCluster| SrcAccessOrder|   SrcVectorDim|      SrcScalar|      DstScalar| AddExtraM|   ThreadCluster|  ThreadCluster| SrcAccessOrder|  SrcVectorDim|      SrcScalar|      DstScalar| AddExtraN| SrcDstVectorDim|       DstScalar|
+        //#############################################################################|       |        |        |        |   Operation|   Operation|   Operation|     Specialization|Spatial|      |      |      |      |   |     |     | Wave| Wave| Lengths_K0_M_K1|   ArrangeOrder|               |               |      PerVector|   PerVector_K1|          | Lengths_K0_N_K1|   ArrangeOrder|               |              |      PerVector|   PerVector_K1|          |                |       PerVector|
+        //#############################################################################|       |        |        |        |            |            |            |                   |       |      |      |      |      |   |     |     |     |     |                |               |               |               |               |               |          |                |               |               |              |               |               |          |                |                |
+        DeviceConvndBwdDataXdl_Input_N_Di_Hi_Wi_C_Weight_K_Z_Y_X_C_Output_N_Do_Ho_Wo_K<    F16,     F16,     F16,     F32, PassThrough, PassThrough, PassThrough, ConvBwdDataDefault,   2,     256,   256,   128,     4,   8,  32,   32,    4,    2,     S<4, 64, 1>,     S<1, 0, 2>,     S<1, 0, 2>,              2,              8,              8,      true,     S<4, 64, 1>,     S<2, 0, 1>,     S<0, 2, 1>,             1,              2,              8,      true,               7,               1>,
+        DeviceConvndBwdDataXdl_Input_N_Di_Hi_Wi_C_Weight_K_Z_Y_X_C_Output_N_Do_Ho_Wo_K<    F16,     F16,     F16,     F32, PassThrough, PassThrough, PassThrough, ConvBwdDataDefault,   2,     256,   128,   256,     4,   8,  32,   32,    2,    4,     S<4, 64, 1>,     S<1, 0, 2>,     S<1, 0, 2>,              2,              8,              8,      true,     S<4, 64, 1>,     S<2, 0, 1>,     S<0, 2, 1>,             1,              4,              8,      true,               7,               1>,
+        DeviceConvndBwdDataXdl_Input_N_Di_Hi_Wi_C_Weight_K_Z_Y_X_C_Output_N_Do_Ho_Wo_K<    F16,     F16,     F16,     F32, PassThrough, PassThrough, PassThrough, ConvBwdDataDefault,   2,     128,   128,   128,     4,   8,  32,   32,    4,    2,     S<4, 32, 1>,     S<1, 0, 2>,     S<1, 0, 2>,              2,              8,              8,      true,     S<4, 32, 1>,     S<2, 0, 1>,     S<0, 2, 1>,             1,              4,              8,      true,               7,               1>,
+#if 1
+        DeviceConvndBwdDataXdl_Input_N_Di_Hi_Wi_C_Weight_K_Z_Y_X_C_Output_N_Do_Ho_Wo_K<    F16,     F16,     F16,     F32, PassThrough, PassThrough, PassThrough, ConvBwdDataDefault,   2,     256,   128,   128,     4,   8,  32,   32,    2,    2,     S<4, 64, 1>,     S<1, 0, 2>,     S<1, 0, 2>,              2,              8,              8,      true,     S<4, 64, 1>,     S<2, 0, 1>,     S<0, 2, 1>,             1,              2,              8,      true,               7,               1>,
+        DeviceConvndBwdDataXdl_Input_N_Di_Hi_Wi_C_Weight_K_Z_Y_X_C_Output_N_Do_Ho_Wo_K<    F16,     F16,     F16,     F32, PassThrough, PassThrough, PassThrough, ConvBwdDataDefault,   2,     256,    64,   128,     4,   8,  32,   32,    1,    2,     S<4, 64, 1>,     S<1, 0, 2>,     S<1, 0, 2>,              2,              8,              8,      true,     S<4, 64, 1>,     S<2, 0, 1>,     S<0, 2, 1>,             1,              2,              8,      true,               7,               1>,
+        DeviceConvndBwdDataXdl_Input_N_Di_Hi_Wi_C_Weight_K_Z_Y_X_C_Output_N_Do_Ho_Wo_K<    F16,     F16,     F16,     F32, PassThrough, PassThrough, PassThrough, ConvBwdDataDefault,   2,     128,    32,   128,     4,   8,  32,   32,    1,    2,     S<4, 32, 1>,     S<1, 0, 2>,     S<1, 0, 2>,              2,              8,              8,      true,     S<4, 32, 1>,     S<2, 0, 1>,     S<0, 2, 1>,             1,              4,              8,      true,               7,               1>,
+        DeviceConvndBwdDataXdl_Input_N_Di_Hi_Wi_C_Weight_K_Z_Y_X_C_Output_N_Do_Ho_Wo_K<    F16,     F16,     F16,     F32, PassThrough, PassThrough, PassThrough, ConvBwdDataDefault,   2,     128,    64,   128,     4,   8,  32,   32,    2,    2,     S<4, 32, 1>,     S<1, 0, 2>,     S<1, 0, 2>,              2,              8,              8,      true,     S<4, 32, 1>,     S<2, 0, 1>,     S<0, 2, 1>,             1,              4,              8,      true,               7,               1>,
+        DeviceConvndBwdDataXdl_Input_N_Di_Hi_Wi_C_Weight_K_Z_Y_X_C_Output_N_Do_Ho_Wo_K<    F16,     F16,     F16,     F32, PassThrough, PassThrough, PassThrough, ConvBwdDataDefault,   2,     256,   128,    64,     4,   8,  32,   32,    2,    1,     S<4, 64, 1>,     S<1, 0, 2>,     S<1, 0, 2>,              2,              8,              8,      true,     S<4, 64, 1>,     S<2, 0, 1>,     S<0, 2, 1>,             1,              1,              8,      true,               7,               1>,
+#endif
+        DeviceConvndBwdDataXdl_Input_N_Di_Hi_Wi_C_Weight_K_Z_Y_X_C_Output_N_Do_Ho_Wo_K<    F16,     F16,     F16,     F32, PassThrough, PassThrough, PassThrough, ConvBwdDataDefault,   2,     128,   128,    64,     4,   8,  32,   32,    2,    2,     S<4, 32, 1>,     S<1, 0, 2>,     S<1, 0, 2>,              2,              8,              8,      true,     S<4, 32, 1>,     S<2, 0, 1>,     S<0, 2, 1>,             1,              2,              8,      true,               7,               1>,       
+        DeviceConvndBwdDataXdl_Input_N_Di_Hi_Wi_C_Weight_K_Z_Y_X_C_Output_N_Do_Ho_Wo_K<    F16,     F16,     F16,     F32, PassThrough, PassThrough, PassThrough, ConvBwdDataDefault,   2,      64,    64,    64,     4,   8,  32,   32,    2,    2,     S<4, 16, 1>,     S<1, 0, 2>,     S<1, 0, 2>,              2,              8,              8,      true,     S<4, 16, 1>,     S<2, 0, 1>,     S<0, 2, 1>,             1,              4,              8,      true,               7,               1>,
+        DeviceConvndBwdDataXdl_Input_N_Di_Hi_Wi_C_Weight_K_Z_Y_X_C_Output_N_Do_Ho_Wo_K<    F16,     F16,     F16,     F32, PassThrough, PassThrough, PassThrough, ConvBwdDataDefault,   2,     128,   128,    32,     4,   8,  32,   32,    2,    1,     S<4, 32, 1>,     S<1, 0, 2>,     S<1, 0, 2>,              2,              8,              8,      true,     S<4, 32, 1>,     S<2, 0, 1>,     S<0, 2, 1>,             1,              1,              8,      true,               7,               1>,
+        DeviceConvndBwdDataXdl_Input_N_Di_Hi_Wi_C_Weight_K_Z_Y_X_C_Output_N_Do_Ho_Wo_K<    F16,     F16,     F16,     F32, PassThrough, PassThrough, PassThrough, ConvBwdDataDefault,   2,      64,    64,    32,     4,   8,  32,   32,    2,    1,     S<4, 16, 1>,     S<1, 0, 2>,     S<1, 0, 2>,              2,              8,              8,      true,     S<4, 16, 1>,     S<2, 0, 1>,     S<0, 2, 1>,             1,              2,              8,      true,               7,               1>,
+        DeviceConvndBwdDataXdl_Input_N_Di_Hi_Wi_C_Weight_K_Z_Y_X_C_Output_N_Do_Ho_Wo_K<    F16,     F16,     F16,     F32, PassThrough, PassThrough, PassThrough, ConvBwdDataDefault,   2,      64,    32,    64,     4,   8,  32,   32,    1,    2,     S<4, 16, 1>,     S<1, 0, 2>,     S<1, 0, 2>,              2,              8,              8,      true,     S<4, 16, 1>,     S<2, 0, 1>,     S<0, 2, 1>,             1,              2,              8,      true,               7,               1>
+        // clang-format on
+        >;
+
+using device_conv2d_bwd_data_xdl_nhwc_kyxc_nhwk_1x1_s1_p0_f16_instances =
+    std::tuple<
+        // clang-format off
+        //#############################################################################| InData| WeiData| OutData| AccData|          In|         Wei|         Out|                     ConvBackward|    Num| Block|  MPer|  NPer| K0Per| K1| MPer| NPer| MXdl| NXdl|  ABlockTransfer| ABlockTransfer| ABlockTransfer| ABlockTransfer| ABlockTransfer| ABlockTransfer| ABlockLds|  BBlockTransfer| BBlockTransfer| BBlockTransfer| BlockTransfer| BBlockTransfer| BBlockTransfer| BBlockLds| CThreadTransfer| CThreadTransfer|
+        //#############################################################################|   Type|    Type|    Type|    Type| Elementwise| Elementwise| Elementwise|                             Data|    Dim|  Size| Block| Block| Block|   |  XDL|  XDL|  Per|  Per|   ThreadCluster|  ThreadCluster| SrcAccessOrder|   SrcVectorDim|      SrcScalar|      DstScalar| AddExtraM|   ThreadCluster|  ThreadCluster| SrcAccessOrder|  SrcVectorDim|      SrcScalar|      DstScalar| AddExtraN| SrcDstVectorDim|       DstScalar|
+        //#############################################################################|       |        |        |        |   Operation|   Operation|   Operation|                   Specialization|Spatial|      |      |      |      |   |     |     | Wave| Wave| Lengths_K0_M_K1|   ArrangeOrder|               |               |      PerVector|   PerVector_K1|          | Lengths_K0_N_K1|   ArrangeOrder|               |              |      PerVector|   PerVector_K1|          |                |       PerVector|
+        //#############################################################################|       |        |        |        |            |            |            |                                 |       |      |      |      |      |   |     |     |     |     |                |               |               |               |               |               |          |                |               |               |              |               |               |          |                |                |
+        DeviceConvndBwdDataXdl_Input_N_Di_Hi_Wi_C_Weight_K_Z_Y_X_C_Output_N_Do_Ho_Wo_K<    F16,     F16,     F16,     F32, PassThrough, PassThrough, PassThrough,  ConvBwdDataFilter1x1Stride1Pad0,   2,     256,   256,   128,     4,  8,   32,   32,    4,    2,     S<4, 64, 1>,     S<1, 0, 2>,     S<1, 0, 2>,              2,              8,              8,      true,     S<4, 64, 1>,     S<2, 0, 1>,     S<0, 2, 1>,             1,              2,              8,      true,               7,               1>,
+        DeviceConvndBwdDataXdl_Input_N_Di_Hi_Wi_C_Weight_K_Z_Y_X_C_Output_N_Do_Ho_Wo_K<    F16,     F16,     F16,     F32, PassThrough, PassThrough, PassThrough,  ConvBwdDataFilter1x1Stride1Pad0,   2,     256,   128,   256,     4,  8,   32,   32,    2,    4,     S<4, 64, 1>,     S<1, 0, 2>,     S<1, 0, 2>,              2,              8,              8,      true,     S<4, 64, 1>,     S<2, 0, 1>,     S<0, 2, 1>,             1,              4,              8,      true,               7,               1>,
+        DeviceConvndBwdDataXdl_Input_N_Di_Hi_Wi_C_Weight_K_Z_Y_X_C_Output_N_Do_Ho_Wo_K<    F16,     F16,     F16,     F32, PassThrough, PassThrough, PassThrough,  ConvBwdDataFilter1x1Stride1Pad0,   2,     128,   128,   128,     4,  8,   32,   32,    4,    2,     S<4, 32, 1>,     S<1, 0, 2>,     S<1, 0, 2>,              2,              8,              8,      true,     S<4, 32, 1>,     S<2, 0, 1>,     S<0, 2, 1>,             1,              4,              8,      true,               7,               1>,
+        DeviceConvndBwdDataXdl_Input_N_Di_Hi_Wi_C_Weight_K_Z_Y_X_C_Output_N_Do_Ho_Wo_K<    F16,     F16,     F16,     F32, PassThrough, PassThrough, PassThrough,  ConvBwdDataFilter1x1Stride1Pad0,   2,     256,   128,   128,     4,  8,   32,   32,    2,    2,     S<4, 64, 1>,     S<1, 0, 2>,     S<1, 0, 2>,              2,              8,              8,      true,     S<4, 64, 1>,     S<2, 0, 1>,     S<0, 2, 1>,             1,              2,              8,      true,               7,               1>,
+        DeviceConvndBwdDataXdl_Input_N_Di_Hi_Wi_C_Weight_K_Z_Y_X_C_Output_N_Do_Ho_Wo_K<    F16,     F16,     F16,     F32, PassThrough, PassThrough, PassThrough,  ConvBwdDataFilter1x1Stride1Pad0,   2,     128,   128,    64,     4,  8,   32,   32,    2,    2,     S<4, 32, 1>,     S<1, 0, 2>,     S<1, 0, 2>,              2,              8,              8,      true,     S<4, 32, 1>,     S<2, 0, 1>,     S<0, 2, 1>,             1,              2,              8,      true,               7,               1>,
+        DeviceConvndBwdDataXdl_Input_N_Di_Hi_Wi_C_Weight_K_Z_Y_X_C_Output_N_Do_Ho_Wo_K<    F16,     F16,     F16,     F32, PassThrough, PassThrough, PassThrough,  ConvBwdDataFilter1x1Stride1Pad0,   2,     128,    64,   128,     4,  8,   32,   32,    2,    2,     S<4, 32, 1>,     S<1, 0, 2>,     S<1, 0, 2>,              2,              8,              8,      true,     S<4, 32, 1>,     S<2, 0, 1>,     S<0, 2, 1>,             1,              4,              8,      true,               7,               1>,
+        DeviceConvndBwdDataXdl_Input_N_Di_Hi_Wi_C_Weight_K_Z_Y_X_C_Output_N_Do_Ho_Wo_K<    F16,     F16,     F16,     F32, PassThrough, PassThrough, PassThrough,  ConvBwdDataFilter1x1Stride1Pad0,   2,      64,    64,    64,     4,  8,   32,   32,    2,    2,     S<4, 16, 1>,     S<1, 0, 2>,     S<1, 0, 2>,              2,              8,              8,      true,     S<4, 16, 1>,     S<2, 0, 1>,     S<0, 2, 1>,             1,              4,              8,      true,               7,               1>,
+        DeviceConvndBwdDataXdl_Input_N_Di_Hi_Wi_C_Weight_K_Z_Y_X_C_Output_N_Do_Ho_Wo_K<    F16,     F16,     F16,     F32, PassThrough, PassThrough, PassThrough,  ConvBwdDataFilter1x1Stride1Pad0,   2,     256,   128,    64,     4,  8,   32,   32,    2,    1,     S<4, 64, 1>,     S<1, 0, 2>,     S<1, 0, 2>,              2,              8,              8,      true,     S<4, 64, 1>,     S<2, 0, 1>,     S<0, 2, 1>,             1,              1,              8,      true,               7,               1>,
+        DeviceConvndBwdDataXdl_Input_N_Di_Hi_Wi_C_Weight_K_Z_Y_X_C_Output_N_Do_Ho_Wo_K<    F16,     F16,     F16,     F32, PassThrough, PassThrough, PassThrough,  ConvBwdDataFilter1x1Stride1Pad0,   2,     256,    64,   128,     4,  8,   32,   32,    1,    2,     S<4, 64, 1>,     S<1, 0, 2>,     S<1, 0, 2>,              2,              8,              8,      true,     S<4, 64, 1>,     S<2, 0, 1>,     S<0, 2, 1>,             1,              2,              8,      true,               7,               1>,
+        DeviceConvndBwdDataXdl_Input_N_Di_Hi_Wi_C_Weight_K_Z_Y_X_C_Output_N_Do_Ho_Wo_K<    F16,     F16,     F16,     F32, PassThrough, PassThrough, PassThrough,  ConvBwdDataFilter1x1Stride1Pad0,   2,     128,   128,    32,     4,  8,   32,   32,    2,    1,     S<4, 32, 1>,     S<1, 0, 2>,     S<1, 0, 2>,              2,              8,              8,      true,     S<4, 32, 1>,     S<2, 0, 1>,     S<0, 2, 1>,             1,              1,              8,      true,               7,               1>,
+        DeviceConvndBwdDataXdl_Input_N_Di_Hi_Wi_C_Weight_K_Z_Y_X_C_Output_N_Do_Ho_Wo_K<    F16,     F16,     F16,     F32, PassThrough, PassThrough, PassThrough,  ConvBwdDataFilter1x1Stride1Pad0,   2,     128,    32,   128,     4,  8,   32,   32,    1,    2,     S<4, 32, 1>,     S<1, 0, 2>,     S<1, 0, 2>,              2,              8,              8,      true,     S<4, 32, 1>,     S<2, 0, 1>,     S<0, 2, 1>,             1,              4,              8,      true,               7,               1>,
+        DeviceConvndBwdDataXdl_Input_N_Di_Hi_Wi_C_Weight_K_Z_Y_X_C_Output_N_Do_Ho_Wo_K<    F16,     F16,     F16,     F32, PassThrough, PassThrough, PassThrough,  ConvBwdDataFilter1x1Stride1Pad0,   2,      64,    64,    32,     4,  8,   32,   32,    2,    1,     S<4, 16, 1>,     S<1, 0, 2>,     S<1, 0, 2>,              2,              8,              8,      true,     S<4, 16, 1>,     S<2, 0, 1>,     S<0, 2, 1>,             1,              2,              8,      true,               7,               1>,
+        DeviceConvndBwdDataXdl_Input_N_Di_Hi_Wi_C_Weight_K_Z_Y_X_C_Output_N_Do_Ho_Wo_K<    F16,     F16,     F16,     F32, PassThrough, PassThrough, PassThrough,  ConvBwdDataFilter1x1Stride1Pad0,   2,      64,    32,    64,     4,  8,   32,   32,    1,    2,     S<4, 16, 1>,     S<1, 0, 2>,     S<1, 0, 2>,              2,              8,              8,      true,     S<4, 16, 1>,     S<2, 0, 1>,     S<0, 2, 1>,             1,              2,              8,      true,               7,               1>
+        // clang-format on
+        >;
+
+void add_device_conv2d_bwd_data_xdl_nhwc_kyxc_nhwk_f16_instances(
+    std::vector<DeviceConvBwdDataPtr<PassThrough, PassThrough, PassThrough>>& instances)
+{
+    add_device_operation_instances(instances,
+                                   device_conv2d_bwd_data_xdl_nhwc_kyxc_nhwk_f16_instances{});
+    add_device_operation_instances(
+        instances, device_conv2d_bwd_data_xdl_nhwc_kyxc_nhwk_1x1_s1_p0_f16_instances{});
+}
+
+} // namespace device_conv2d_bwd_data_instance
+} // namespace device
+} // namespace tensor_operation
+} // namespace ck

library/src/tensor_operation_instance/gpu/convnd_bwd_data/device_conv2d_bwd_data_xdl_nhwc_kyxc_nhwk_f32_instance.cpp

+#include <stdlib.h>
+#include "config.hpp"
+#include "device_convnd_bwd_data_xdl_ndhwc_kzyxc_ndhwk.hpp"
+#include "element_wise_operation.hpp"
+#include "device_operation_instance.hpp"
+
+namespace ck {
+namespace tensor_operation {
+namespace device {
+namespace device_conv2d_bwd_data_instance {
+
+using F32 = float;
+
+template <ck::index_t... Is>
+using S = ck::Sequence<Is...>;
+
+using PassThrough = ck::tensor_operation::element_wise::PassThrough;
+static constexpr auto ConvBwdDataDefault =
+    ck::tensor_operation::device::ConvolutionBackwardDataSpecialization_t::Default;
+
+static constexpr auto ConvBwdDataFilter1x1Stride1Pad0 =
+    ck::tensor_operation::device::ConvolutionBackwardDataSpecialization_t::Filter1x1Stride1Pad0;
+
+// Compilation parameters for in[n, hi, wi, c] * wei[k, y, x, c] = out[n, ho, wo, k]
+using device_conv2d_bwd_data_xdl_nhwc_kyxc_nhwk_f32_instances =
+    std::tuple<
+        // clang-format off
+        //#############################################################################| InData| WeiData| OutData| AccData|          In|         Wei|         Out|        ConvBackward|    Num| Block|  MPer|  NPer| K0Per| K1| MPer| NPer| MXdl| NXdl|  ABlockTransfer| ABlockTransfer| ABlockTransfer| ABlockTransfer| ABlockTransfer| ABlockTransfer| ABlockLds|  BBlockTransfer| BBlockTransfer| BBlockTransfer| BlockTransfer| BBlockTransfer| BBlockTransfer| BBlockLds| CThreadTransfer| CThreadTransfer|
+        //#############################################################################|   Type|    Type|    Type|    Type| Elementwise| Elementwise| Elementwise|                Data|    Dim|  Size| Block| Block| Block|   |  XDL|  XDL|  Per|  Per|   ThreadCluster|  ThreadCluster| SrcAccessOrder|   SrcVectorDim|      SrcScalar|      DstScalar| AddExtraM|   ThreadCluster|  ThreadCluster| SrcAccessOrder|  SrcVectorDim|      SrcScalar|      DstScalar| AddExtraN| SrcDstVectorDim|       DstScalar|
+        //#############################################################################|       |        |        |        |   Operation|   Operation|   Operation|      Specialization|Spatial|      |      |      |      |   |     |     | Wave| Wave| Lengths_K0_M_K1|   ArrangeOrder|               |               |      PerVector|   PerVector_K1|          | Lengths_K0_N_K1|   ArrangeOrder|               |              |      PerVector|   PerVector_K1|          |                |       PerVector|
+        //#############################################################################|       |        |        |        |            |            |            |                    |       |      |      |      |      |   |     |     |     |     |                |               |               |               |               |               |          |                |               |               |              |               |               |          |                |                |
+        DeviceConvndBwdDataXdl_Input_N_Di_Hi_Wi_C_Weight_K_Z_Y_X_C_Output_N_Do_Ho_Wo_K<    F32,     F32,     F32,     F32, PassThrough, PassThrough, PassThrough,  ConvBwdDataDefault,   2,     256,   256,   128,     4,  4,   32,   32,    4,    2,     S<4, 64, 1>,     S<1, 0, 2>,     S<1, 0, 2>,              2,              4,              4,      true,     S<4, 64, 1>,     S<2, 0, 1>,     S<0, 2, 1>,             1,              2,              4,      true,               7,               1>,
+        DeviceConvndBwdDataXdl_Input_N_Di_Hi_Wi_C_Weight_K_Z_Y_X_C_Output_N_Do_Ho_Wo_K<    F32,     F32,     F32,     F32, PassThrough, PassThrough, PassThrough,  ConvBwdDataDefault,   2,     256,   128,   256,     4,  4,   32,   32,    2,    4,     S<4, 64, 1>,     S<1, 0, 2>,     S<1, 0, 2>,              2,              4,              4,      true,     S<4, 64, 1>,     S<2, 0, 1>,     S<0, 2, 1>,             1,              4,              4,      true,               7,               1>,
+        DeviceConvndBwdDataXdl_Input_N_Di_Hi_Wi_C_Weight_K_Z_Y_X_C_Output_N_Do_Ho_Wo_K<    F32,     F32,     F32,     F32, PassThrough, PassThrough, PassThrough,  ConvBwdDataDefault,   2,     128,   128,   128,     4,  4,   32,   32,    4,    2,     S<4, 32, 1>,     S<1, 0, 2>,     S<1, 0, 2>,              2,              4,              4,      true,     S<4, 32, 1>,     S<2, 0, 1>,     S<0, 2, 1>,             1,              4,              4,      true,               7,               1>,
+        DeviceConvndBwdDataXdl_Input_N_Di_Hi_Wi_C_Weight_K_Z_Y_X_C_Output_N_Do_Ho_Wo_K<    F32,     F32,     F32,     F32, PassThrough, PassThrough, PassThrough,  ConvBwdDataDefault,   2,     256,   128,   128,     4,  4,   32,   32,    2,    2,     S<4, 64, 1>,     S<1, 0, 2>,     S<1, 0, 2>,              2,              4,              4,      true,     S<4, 64, 1>,     S<2, 0, 1>,     S<0, 2, 1>,             1,              2,              4,      true,               7,               1>,
+        DeviceConvndBwdDataXdl_Input_N_Di_Hi_Wi_C_Weight_K_Z_Y_X_C_Output_N_Do_Ho_Wo_K<    F32,     F32,     F32,     F32, PassThrough, PassThrough, PassThrough,  ConvBwdDataDefault,   2,     128,   128,    64,     4,  4,   32,   32,    2,    2,     S<4, 32, 1>,     S<1, 0, 2>,     S<1, 0, 2>,              2,              4,              4,      true,     S<4, 32, 1>,     S<2, 0, 1>,     S<0, 2, 1>,             1,              2,              4,      true,               7,               1>,
+        DeviceConvndBwdDataXdl_Input_N_Di_Hi_Wi_C_Weight_K_Z_Y_X_C_Output_N_Do_Ho_Wo_K<    F32,     F32,     F32,     F32, PassThrough, PassThrough, PassThrough,  ConvBwdDataDefault,   2,     128,    64,   128,     4,  4,   32,   32,    2,    2,     S<4, 32, 1>,     S<1, 0, 2>,     S<1, 0, 2>,              2,              4,              4,      true,     S<4, 32, 1>,     S<2, 0, 1>,     S<0, 2, 1>,             1,              4,              4,      true,               7,               1>,
+        DeviceConvndBwdDataXdl_Input_N_Di_Hi_Wi_C_Weight_K_Z_Y_X_C_Output_N_Do_Ho_Wo_K<    F32,     F32,     F32,     F32, PassThrough, PassThrough, PassThrough,  ConvBwdDataDefault,   2,      64,    64,    64,     4,  4,   32,   32,    2,    2,     S<4, 16, 1>,     S<1, 0, 2>,     S<1, 0, 2>,              2,              4,              4,      true,     S<4, 16, 1>,     S<2, 0, 1>,     S<0, 2, 1>,             1,              4,              4,      true,               7,               1>,
+        DeviceConvndBwdDataXdl_Input_N_Di_Hi_Wi_C_Weight_K_Z_Y_X_C_Output_N_Do_Ho_Wo_K<    F32,     F32,     F32,     F32, PassThrough, PassThrough, PassThrough,  ConvBwdDataDefault,   2,     256,   128,    64,     4,  4,   32,   32,    2,    1,     S<4, 64, 1>,     S<1, 0, 2>,     S<1, 0, 2>,              2,              4,              4,      true,     S<4, 64, 1>,     S<2, 0, 1>,     S<0, 2, 1>,             1,              1,              4,      true,               7,               1>,
+        DeviceConvndBwdDataXdl_Input_N_Di_Hi_Wi_C_Weight_K_Z_Y_X_C_Output_N_Do_Ho_Wo_K<    F32,     F32,     F32,     F32, PassThrough, PassThrough, PassThrough,  ConvBwdDataDefault,   2,     256,    64,   128,     4,  4,   32,   32,    1,    2,     S<4, 64, 1>,     S<1, 0, 2>,     S<1, 0, 2>,              2,              4,              4,      true,     S<4, 64, 1>,     S<2, 0, 1>,     S<0, 2, 1>,             1,              2,              4,      true,               7,               1>,
+        DeviceConvndBwdDataXdl_Input_N_Di_Hi_Wi_C_Weight_K_Z_Y_X_C_Output_N_Do_Ho_Wo_K<    F32,     F32,     F32,     F32, PassThrough, PassThrough, PassThrough,  ConvBwdDataDefault,   2,     128,   128,    32,     4,  4,   32,   32,    2,    1,     S<4, 32, 1>,     S<1, 0, 2>,     S<1, 0, 2>,              2,              4,              4,      true,     S<4, 32, 1>,     S<2, 0, 1>,     S<0, 2, 1>,             1,              1,              4,      true,               7,               1>,
+        DeviceConvndBwdDataXdl_Input_N_Di_Hi_Wi_C_Weight_K_Z_Y_X_C_Output_N_Do_Ho_Wo_K<    F32,     F32,     F32,     F32, PassThrough, PassThrough, PassThrough,  ConvBwdDataDefault,   2,     128,    32,   128,     4,  4,   32,   32,    1,    2,     S<4, 32, 1>,     S<1, 0, 2>,     S<1, 0, 2>,              2,              4,              4,      true,     S<4, 32, 1>,     S<2, 0, 1>,     S<0, 2, 1>,             1,              4,              4,      true,               7,               1>,
+        DeviceConvndBwdDataXdl_Input_N_Di_Hi_Wi_C_Weight_K_Z_Y_X_C_Output_N_Do_Ho_Wo_K<    F32,     F32,     F32,     F32, PassThrough, PassThrough, PassThrough,  ConvBwdDataDefault,   2,      64,    64,    32,     4,  4,   32,   32,    2,    1,     S<4, 16, 1>,     S<1, 0, 2>,     S<1, 0, 2>,              2,              4,              4,      true,     S<4, 16, 1>,     S<2, 0, 1>,     S<0, 2, 1>,             1,              2,              4,      true,               7,               1>,
+        DeviceConvndBwdDataXdl_Input_N_Di_Hi_Wi_C_Weight_K_Z_Y_X_C_Output_N_Do_Ho_Wo_K<    F32,     F32,     F32,     F32, PassThrough, PassThrough, PassThrough,  ConvBwdDataDefault,   2,      64,    32,    64,     4,  4,   32,   32,    1,    2,     S<4, 16, 1>,     S<1, 0, 2>,     S<1, 0, 2>,              2,              4,              4,      true,     S<4, 16, 1>,     S<2, 0, 1>,     S<0, 2, 1>,             1,              4,              4,      true,               7,               1>
+        // clang-format on
+        >;
+
+using device_conv2d_bwd_data_xdl_nhwc_kyxc_nhwk_1x1_s1_p0_f32_instances =
+    std::tuple<
+        // clang-format off
+        //#############################################################################| InData| WeiData| OutData| AccData|          In|         Wei|         Out|                     ConvBackward|    Num| Block|  MPer|  NPer| K0Per| K1| MPer| NPer| MXdl| NXdl|  ABlockTransfer| ABlockTransfer| ABlockTransfer| ABlockTransfer| ABlockTransfer| ABlockTransfer| ABlockLds|  BBlockTransfer| BBlockTransfer| BBlockTransfer| BlockTransfer| BBlockTransfer| BBlockTransfer| BBlockLds| CThreadTransfer| CThreadTransfer|
+        //#############################################################################|   Type|    Type|    Type|    Type| Elementwise| Elementwise| Elementwise|                             Data|    Dim|  Size| Block| Block| Block|   |  XDL|  XDL|  Per|  Per|   ThreadCluster|  ThreadCluster| SrcAccessOrder|   SrcVectorDim|      SrcScalar|      DstScalar| AddExtraM|   ThreadCluster|  ThreadCluster| SrcAccessOrder|  SrcVectorDim|      SrcScalar|      DstScalar| AddExtraN| SrcDstVectorDim|       DstScalar|
+        //#############################################################################|       |        |        |        |   Operation|   Operation|   Operation|                   Specialization|Spatial|      |      |      |      |   |     |     | Wave| Wave| Lengths_K0_M_K1|   ArrangeOrder|               |               |      PerVector|   PerVector_K1|          | Lengths_K0_N_K1|   ArrangeOrder|               |              |      PerVector|   PerVector_K1|          |                |       PerVector|
+        //#############################################################################|       |        |        |        |            |            |            |                                 |       |      |      |      |      |   |     |     |     |     |                |               |               |               |               |               |          |                |               |               |              |               |               |          |                |                |
+        DeviceConvndBwdDataXdl_Input_N_Di_Hi_Wi_C_Weight_K_Z_Y_X_C_Output_N_Do_Ho_Wo_K<    F32,     F32,     F32,     F32, PassThrough, PassThrough, PassThrough,  ConvBwdDataFilter1x1Stride1Pad0,   2,     256,   256,   128,     4,  4,   32,   32,    4,    2,     S<4, 64, 1>,     S<1, 0, 2>,     S<1, 0, 2>,              2,              4,              4,      true,     S<4, 64, 1>,     S<2, 0, 1>,     S<0, 2, 1>,             1,              2,              4,      true,               7,               1>,
+        DeviceConvndBwdDataXdl_Input_N_Di_Hi_Wi_C_Weight_K_Z_Y_X_C_Output_N_Do_Ho_Wo_K<    F32,     F32,     F32,     F32, PassThrough, PassThrough, PassThrough,  ConvBwdDataFilter1x1Stride1Pad0,   2,     256,   128,   256,     4,  4,   32,   32,    2,    4,     S<4, 64, 1>,     S<1, 0, 2>,     S<1, 0, 2>,              2,              4,              4,      true,     S<4, 64, 1>,     S<2, 0, 1>,     S<0, 2, 1>,             1,              4,              4,      true,               7,               1>,
+        DeviceConvndBwdDataXdl_Input_N_Di_Hi_Wi_C_Weight_K_Z_Y_X_C_Output_N_Do_Ho_Wo_K<    F32,     F32,     F32,     F32, PassThrough, PassThrough, PassThrough,  ConvBwdDataFilter1x1Stride1Pad0,   2,     128,   128,   128,     4,  4,   32,   32,    4,    2,     S<4, 32, 1>,     S<1, 0, 2>,     S<1, 0, 2>,              2,              4,              4,      true,     S<4, 32, 1>,     S<2, 0, 1>,     S<0, 2, 1>,             1,              4,              4,      true,               7,               1>,
+        DeviceConvndBwdDataXdl_Input_N_Di_Hi_Wi_C_Weight_K_Z_Y_X_C_Output_N_Do_Ho_Wo_K<    F32,     F32,     F32,     F32, PassThrough, PassThrough, PassThrough,  ConvBwdDataFilter1x1Stride1Pad0,   2,     256,   128,   128,     4,  4,   32,   32,    2,    2,     S<4, 64, 1>,     S<1, 0, 2>,     S<1, 0, 2>,              2,              4,              4,      true,     S<4, 64, 1>,     S<2, 0, 1>,     S<0, 2, 1>,             1,              2,              4,      true,               7,               1>,
+        DeviceConvndBwdDataXdl_Input_N_Di_Hi_Wi_C_Weight_K_Z_Y_X_C_Output_N_Do_Ho_Wo_K<    F32,     F32,     F32,     F32, PassThrough, PassThrough, PassThrough,  ConvBwdDataFilter1x1Stride1Pad0,   2,     128,   128,    64,     4,  4,   32,   32,    2,    2,     S<4, 32, 1>,     S<1, 0, 2>,     S<1, 0, 2>,              2,              4,              4,      true,     S<4, 32, 1>,     S<2, 0, 1>,     S<0, 2, 1>,             1,              2,              4,      true,               7,               1>,
+        DeviceConvndBwdDataXdl_Input_N_Di_Hi_Wi_C_Weight_K_Z_Y_X_C_Output_N_Do_Ho_Wo_K<    F32,     F32,     F32,     F32, PassThrough, PassThrough, PassThrough,  ConvBwdDataFilter1x1Stride1Pad0,   2,     128,    64,   128,     4,  4,   32,   32,    2,    2,     S<4, 32, 1>,     S<1, 0, 2>,     S<1, 0, 2>,              2,              4,              4,      true,     S<4, 32, 1>,     S<2, 0, 1>,     S<0, 2, 1>,             1,              4,              4,      true,               7,               1>,
+        DeviceConvndBwdDataXdl_Input_N_Di_Hi_Wi_C_Weight_K_Z_Y_X_C_Output_N_Do_Ho_Wo_K<    F32,     F32,     F32,     F32, PassThrough, PassThrough, PassThrough,  ConvBwdDataFilter1x1Stride1Pad0,   2,      64,    64,    64,     4,  4,   32,   32,    2,    2,     S<4, 16, 1>,     S<1, 0, 2>,     S<1, 0, 2>,              2,              4,              4,      true,     S<4, 16, 1>,     S<2, 0, 1>,     S<0, 2, 1>,             1,              4,              4,      true,               7,               1>,
+        DeviceConvndBwdDataXdl_Input_N_Di_Hi_Wi_C_Weight_K_Z_Y_X_C_Output_N_Do_Ho_Wo_K<    F32,     F32,     F32,     F32, PassThrough, PassThrough, PassThrough,  ConvBwdDataFilter1x1Stride1Pad0,   2,     256,   128,    64,     4,  4,   32,   32,    2,    1,     S<4, 64, 1>,     S<1, 0, 2>,     S<1, 0, 2>,              2,              4,              4,      true,     S<4, 64, 1>,     S<2, 0, 1>,     S<0, 2, 1>,             1,              1,              4,      true,               7,               1>,
+        DeviceConvndBwdDataXdl_Input_N_Di_Hi_Wi_C_Weight_K_Z_Y_X_C_Output_N_Do_Ho_Wo_K<    F32,     F32,     F32,     F32, PassThrough, PassThrough, PassThrough,  ConvBwdDataFilter1x1Stride1Pad0,   2,     256,    64,   128,     4,  4,   32,   32,    1,    2,     S<4, 64, 1>,     S<1, 0, 2>,     S<1, 0, 2>,              2,              4,              4,      true,     S<4, 64, 1>,     S<2, 0, 1>,     S<0, 2, 1>,             1,              2,              4,      true,               7,               1>,
+        DeviceConvndBwdDataXdl_Input_N_Di_Hi_Wi_C_Weight_K_Z_Y_X_C_Output_N_Do_Ho_Wo_K<    F32,     F32,     F32,     F32, PassThrough, PassThrough, PassThrough,  ConvBwdDataFilter1x1Stride1Pad0,   2,     128,   128,    32,     4,  4,   32,   32,    2,    1,     S<4, 32, 1>,     S<1, 0, 2>,     S<1, 0, 2>,              2,              4,              4,      true,     S<4, 32, 1>,     S<2, 0, 1>,     S<0, 2, 1>,             1,              1,              4,      true,               7,               1>,
+        DeviceConvndBwdDataXdl_Input_N_Di_Hi_Wi_C_Weight_K_Z_Y_X_C_Output_N_Do_Ho_Wo_K<    F32,     F32,     F32,     F32, PassThrough, PassThrough, PassThrough,  ConvBwdDataFilter1x1Stride1Pad0,   2,     128,    32,   128,     4,  4,   32,   32,    1,    2,     S<4, 32, 1>,     S<1, 0, 2>,     S<1, 0, 2>,              2,              4,              4,      true,     S<4, 32, 1>,     S<2, 0, 1>,     S<0, 2, 1>,             1,              4,              4,      true,               7,               1>,
+        DeviceConvndBwdDataXdl_Input_N_Di_Hi_Wi_C_Weight_K_Z_Y_X_C_Output_N_Do_Ho_Wo_K<    F32,     F32,     F32,     F32, PassThrough, PassThrough, PassThrough,  ConvBwdDataFilter1x1Stride1Pad0,   2,      64,    64,    32,     4,  4,   32,   32,    2,    1,     S<4, 16, 1>,     S<1, 0, 2>,     S<1, 0, 2>,              2,              4,              4,      true,     S<4, 16, 1>,     S<2, 0, 1>,     S<0, 2, 1>,             1,              2,              4,      true,               7,               1>,
+        DeviceConvndBwdDataXdl_Input_N_Di_Hi_Wi_C_Weight_K_Z_Y_X_C_Output_N_Do_Ho_Wo_K<    F32,     F32,     F32,     F32, PassThrough, PassThrough, PassThrough,  ConvBwdDataFilter1x1Stride1Pad0,   2,      64,    32,    64,     4,  4,   32,   32,    1,    2,     S<4, 16, 1>,     S<1, 0, 2>,     S<1, 0, 2>,              2,              4,              4,      true,     S<4, 16, 1>,     S<2, 0, 1>,     S<0, 2, 1>,             1,              4,              4,      true,               7,               1>
+        // clang-format on
+        >;
+
+void add_device_conv2d_bwd_data_xdl_nhwc_kyxc_nhwk_f32_instances(
+    std::vector<DeviceConvBwdDataPtr<PassThrough, PassThrough, PassThrough>>& instances)
+{
+    add_device_operation_instances(instances,
+                                   device_conv2d_bwd_data_xdl_nhwc_kyxc_nhwk_f32_instances{});
+    add_device_operation_instances(
+        instances, device_conv2d_bwd_data_xdl_nhwc_kyxc_nhwk_1x1_s1_p0_f32_instances{});
+}
+
+} // namespace device_conv2d_bwd_data_instance
+} // namespace device
+} // namespace tensor_operation
+} // namespace ck

library/src/tensor_operation_instance/gpu/convnd_bwd_data/device_conv2d_bwd_data_xdl_nhwc_kyxc_nhwk_int8_instance.cpp

+#include <stdlib.h>
+#include "config.hpp"
+#include "device_convnd_bwd_data_xdl_ndhwc_kzyxc_ndhwk.hpp"
+#include "element_wise_operation.hpp"
+#include "device_operation_instance.hpp"
+
+namespace ck {
+namespace tensor_operation {
+namespace device {
+namespace device_conv2d_bwd_data_instance {
+
+using DataType = int8_t;
+using AccType  = int32_t;
+
+template <ck::index_t... Is>
+using S = ck::Sequence<Is...>;
+
+using PassThrough = ck::tensor_operation::element_wise::PassThrough;
+static constexpr auto ConvBwdDataDefault =
+    ck::tensor_operation::device::ConvolutionBackwardDataSpecialization_t::Default;
+
+static constexpr auto ConvBwdDataFilter1x1Stride1Pad0 =
+    ck::tensor_operation::device::ConvolutionBackwardDataSpecialization_t::Filter1x1Stride1Pad0;
+
+// Compilation parameters for in[n, hi, wi, c] * wei[k, y, x, c] = out[n, ho, wo, k]
+using device_conv2d_bwd_data_xdl_nhwc_kyxc_nhwk_int8_instances =
+    std::tuple<
+        // clang-format off
+        //#############################################################################|    InData|   WeiData|   OutData|    AccData|          In|         Wei|         Out|        ConvBackward|    Num| Block|  MPer|  NPer| K0Per|  K1| MPer| NPer| MXdl| NXdl|  ABlockTransfer| ABlockTransfer| ABlockTransfer| ABlockTransfer| ABlockTransfer| ABlockTransfer| ABlockLds|  BBlockTransfer| BBlockTransfer| BBlockTransfer| BlockTransfer| BBlockTransfer| BBlockTransfer| BBlockLds| CThreadTransfer| CThreadTransfer|
+        //#############################################################################|      Type|      Type|      Type|       Type| Elementwise| Elementwise| Elementwise|                Data|    Dim|  Size| Block| Block| Block|    |  XDL|  XDL|  Per|  Per|   ThreadCluster|  ThreadCluster| SrcAccessOrder|   SrcVectorDim|      SrcScalar|      DstScalar| AddExtraM|   ThreadCluster|  ThreadCluster| SrcAccessOrder|  SrcVectorDim|      SrcScalar|      DstScalar| AddExtraN| SrcDstVectorDim|       DstScalar|
+        //#############################################################################|          |          |          |           |   Operation|   Operation|   Operation|      Specialization|Spatial|      |      |      |      |    |     |     | Wave| Wave| Lengths_K0_M_K1|   ArrangeOrder|               |               |      PerVector|   PerVector_K1|          | Lengths_K0_N_K1|   ArrangeOrder|               |              |      PerVector|   PerVector_K1|          |                |       PerVector|
+        //#############################################################################|          |          |          |           |            |            |            |                    |       |      |      |      |      |    |     |     |     |     |                |               |               |               |               |               |          |                |               |               |              |               |               |          |                |                |
+        DeviceConvndBwdDataXdl_Input_N_Di_Hi_Wi_C_Weight_K_Z_Y_X_C_Output_N_Do_Ho_Wo_K<  DataType,  DataType,  DataType,    AccType, PassThrough, PassThrough, PassThrough,  ConvBwdDataDefault,   2,     256,   128,   256,     4,  16,   32,   32,    2,    4,     S<4, 64, 1>,     S<1, 0, 2>,     S<1, 0, 2>,              2,             16,             16,      true,     S<4, 64, 1>,     S<2, 0, 1>,     S<0, 2, 1>,             1,              4,             16,      true,               7,               1>,
+        DeviceConvndBwdDataXdl_Input_N_Di_Hi_Wi_C_Weight_K_Z_Y_X_C_Output_N_Do_Ho_Wo_K<  DataType,  DataType,  DataType,    AccType, PassThrough, PassThrough, PassThrough,  ConvBwdDataDefault,   2,     128,   128,   128,     4,  16,   32,   32,    4,    2,     S<4, 32, 1>,     S<1, 0, 2>,     S<1, 0, 2>,              2,             16,             16,      true,     S<4, 32, 1>,     S<2, 0, 1>,     S<0, 2, 1>,             1,              2,             16,      true,               7,               1>,
+    #if 1
+        DeviceConvndBwdDataXdl_Input_N_Di_Hi_Wi_C_Weight_K_Z_Y_X_C_Output_N_Do_Ho_Wo_K<  DataType,  DataType,  DataType,    AccType, PassThrough, PassThrough, PassThrough,  ConvBwdDataDefault,   2,     256,   256,   128,     4,  16,   32,   32,    4,    2,     S<4, 64, 1>,     S<1, 0, 2>,     S<1, 0, 2>,              2,             16,             16,      true,     S<4, 64, 1>,     S<2, 0, 1>,     S<0, 2, 1>,             1,              2,             16,      true,               7,               1>,
+        DeviceConvndBwdDataXdl_Input_N_Di_Hi_Wi_C_Weight_K_Z_Y_X_C_Output_N_Do_Ho_Wo_K<  DataType,  DataType,  DataType,    AccType, PassThrough, PassThrough, PassThrough,  ConvBwdDataDefault,   2,     256,   128,   128,     4,  16,   32,   32,    2,    2,     S<4, 64, 1>,     S<1, 0, 2>,     S<1, 0, 2>,              2,             16,             16,      true,     S<4, 64, 1>,     S<2, 0, 1>,     S<0, 2, 1>,             1,              2,             16,      true,               7,               1>,
+    #endif
+        DeviceConvndBwdDataXdl_Input_N_Di_Hi_Wi_C_Weight_K_Z_Y_X_C_Output_N_Do_Ho_Wo_K<  DataType,  DataType,  DataType,    AccType, PassThrough, PassThrough, PassThrough,  ConvBwdDataDefault,   2,     128,   128,    64,     4,  16,   32,   32,    2,    2,     S<4, 32, 1>,     S<1, 0, 2>,     S<1, 0, 2>,              2,             16,             16,      true,     S<4, 32, 1>,     S<2, 0, 1>,     S<0, 2, 1>,             1,              2,             16,      true,               7,               1>,
+        DeviceConvndBwdDataXdl_Input_N_Di_Hi_Wi_C_Weight_K_Z_Y_X_C_Output_N_Do_Ho_Wo_K<  DataType,  DataType,  DataType,    AccType, PassThrough, PassThrough, PassThrough,  ConvBwdDataDefault,   2,     128,    64,   128,     4,  16,   32,   32,    2,    2,     S<4, 32, 1>,     S<1, 0, 2>,     S<1, 0, 2>,              2,             16,             16,      true,     S<4, 32, 1>,     S<2, 0, 1>,     S<0, 2, 1>,             1,              4,             16,      true,               7,               1>,
+        DeviceConvndBwdDataXdl_Input_N_Di_Hi_Wi_C_Weight_K_Z_Y_X_C_Output_N_Do_Ho_Wo_K<  DataType,  DataType,  DataType,    AccType, PassThrough, PassThrough, PassThrough,  ConvBwdDataDefault,   2,      64,    64,    64,     4,  16,   32,   32,    2,    2,     S<4, 16, 1>,     S<1, 0, 2>,     S<1, 0, 2>,              2,             16,             16,      true,     S<4, 16, 1>,     S<2, 0, 1>,     S<0, 2, 1>,             1,              4,             16,      true,               7,               1>,
+        DeviceConvndBwdDataXdl_Input_N_Di_Hi_Wi_C_Weight_K_Z_Y_X_C_Output_N_Do_Ho_Wo_K<  DataType,  DataType,  DataType,    AccType, PassThrough, PassThrough, PassThrough,  ConvBwdDataDefault,   2,     256,   128,    64,     4,  16,   32,   32,    2,    1,     S<4, 64, 1>,     S<1, 0, 2>,     S<1, 0, 2>,              2,             16,             16,      true,     S<4, 64, 1>,     S<2, 0, 1>,     S<0, 2, 1>,             1,              1,             16,      true,               7,               1>,
+        DeviceConvndBwdDataXdl_Input_N_Di_Hi_Wi_C_Weight_K_Z_Y_X_C_Output_N_Do_Ho_Wo_K<  DataType,  DataType,  DataType,    AccType, PassThrough, PassThrough, PassThrough,  ConvBwdDataDefault,   2,     256,    64,   128,     4,  16,   32,   32,    1,    2,     S<4, 64, 1>,     S<1, 0, 2>,     S<1, 0, 2>,              2,             16,             16,      true,     S<4, 64, 1>,     S<2, 0, 1>,     S<0, 2, 1>,             1,              2,             16,      true,               7,               1>,
+        DeviceConvndBwdDataXdl_Input_N_Di_Hi_Wi_C_Weight_K_Z_Y_X_C_Output_N_Do_Ho_Wo_K<  DataType,  DataType,  DataType,    AccType, PassThrough, PassThrough, PassThrough,  ConvBwdDataDefault,   2,     128,   128,    32,     4,  16,   32,   32,    2,    1,     S<4, 32, 1>,     S<1, 0, 2>,     S<1, 0, 2>,              2,             16,             16,      true,     S<4, 32, 1>,     S<2, 0, 1>,     S<0, 2, 1>,             1,              1,             16,      true,               7,               1>,
+        DeviceConvndBwdDataXdl_Input_N_Di_Hi_Wi_C_Weight_K_Z_Y_X_C_Output_N_Do_Ho_Wo_K<  DataType,  DataType,  DataType,    AccType, PassThrough, PassThrough, PassThrough,  ConvBwdDataDefault,   2,     128,    32,   128,     4,  16,   32,   32,    1,    2,     S<4, 32, 1>,     S<1, 0, 2>,     S<1, 0, 2>,              2,             16,             16,      true,     S<4, 32, 1>,     S<2, 0, 1>,     S<0, 2, 1>,             1,              4,             16,      true,               7,               1>,
+        DeviceConvndBwdDataXdl_Input_N_Di_Hi_Wi_C_Weight_K_Z_Y_X_C_Output_N_Do_Ho_Wo_K<  DataType,  DataType,  DataType,    AccType, PassThrough, PassThrough, PassThrough,  ConvBwdDataDefault,   2,      64,    64,    32,     4,  16,   32,   32,    2,    1,     S<4, 16, 1>,     S<1, 0, 2>,     S<1, 0, 2>,              2,             16,             16,      true,     S<4, 16, 1>,     S<2, 0, 1>,     S<0, 2, 1>,             1,              2,             16,      true,               7,               1>,
+        DeviceConvndBwdDataXdl_Input_N_Di_Hi_Wi_C_Weight_K_Z_Y_X_C_Output_N_Do_Ho_Wo_K<  DataType,  DataType,  DataType,    AccType, PassThrough, PassThrough, PassThrough,  ConvBwdDataDefault,   2,      64,    32,    64,     4,  16,   32,   32,    1,    2,     S<4, 16, 1>,     S<1, 0, 2>,     S<1, 0, 2>,              2,             16,             16,      true,     S<4, 16, 1>,     S<2, 0, 1>,     S<0, 2, 1>,             1,              4,             16,      true,               7,               1>
+        // clang-format on
+        >;
+
+using device_conv2d_bwd_data_xdl_nhwc_kyxc_nhwk_1x1_s1_p0_int8_instances =
+    std::tuple<
+        // clang-format off
+        //##############################################################################|    InData|   WeiData|   OutData|    AccData|          In|         Wei|         Out|                     ConvBackward|    Num| Block|  MPer|  NPer| K0Per|  K1| MPer| NPer| MXdl| NXdl|   ABlockTransfer| ABlockTransfer| ABlockTransfer| ABlockTransfer| ABlockTransfer| ABlockTransfer| ABlockLds|  BBlockTransfer| BBlockTransfer| BBlockTransfer| BlockTransfer| BBlockTransfer| BBlockTransfer| BBlockLds| CThreadTransfer| CThreadTransfer|
+        //##############################################################################|      Type|      Type|      Type|       Type| Elementwise| Elementwise| Elementwise|                             Data|    Dim|  Size| Block| Block| Block|    |  XDL|  XDL|  Per|  Per|    ThreadCluster|  ThreadCluster| SrcAccessOrder|   SrcVectorDim|      SrcScalar|      DstScalar| AddExtraM|   ThreadCluster|  ThreadCluster| SrcAccessOrder|  SrcVectorDim|      SrcScalar|      DstScalar| AddExtraN| SrcDstVectorDim|       DstScalar|
+        //##############################################################################|          |          |          |           |   Operation|   Operation|   Operation|                   Specialization|Spatial|      |      |      |      |    |     |     | Wave| Wave|  Lengths_K0_M_K1|   ArrangeOrder|               |               |      PerVector|   PerVector_K1|          | Lengths_K0_N_K1|   ArrangeOrder|               |              |      PerVector|   PerVector_K1|          |                |       PerVector|
+        //##############################################################################|          |          |          |           |            |            |            |                                 |       |      |      |      |      |    |     |     |     |     |                 |               |               |               |               |               |          |                |               |               |              |               |               |          |                |                |
+        DeviceConvndBwdDataXdl_Input_N_Di_Hi_Wi_C_Weight_K_Z_Y_X_C_Output_N_Do_Ho_Wo_K<   DataType,  DataType,  DataType,    AccType, PassThrough, PassThrough, PassThrough,  ConvBwdDataFilter1x1Stride1Pad0,   2,     256,   256,   128,     4,  16,   32,   32,    4,    2,      S<4, 64, 1>,     S<1, 0, 2>,     S<1, 0, 2>,              2,             16,             16,      true,     S<4, 64, 1>,     S<2, 0, 1>,     S<0, 2, 1>,             1,              2,             16,      true,               7,               1>,
+        DeviceConvndBwdDataXdl_Input_N_Di_Hi_Wi_C_Weight_K_Z_Y_X_C_Output_N_Do_Ho_Wo_K<   DataType,  DataType,  DataType,    AccType, PassThrough, PassThrough, PassThrough,  ConvBwdDataFilter1x1Stride1Pad0,   2,     256,   128,   256,     4,  16,   32,   32,    2,    4,      S<4, 64, 1>,     S<1, 0, 2>,     S<1, 0, 2>,              2,             16,             16,      true,     S<4, 64, 1>,     S<2, 0, 1>,     S<0, 2, 1>,             1,              4,             16,      true,               7,               1>,
+        DeviceConvndBwdDataXdl_Input_N_Di_Hi_Wi_C_Weight_K_Z_Y_X_C_Output_N_Do_Ho_Wo_K<   DataType,  DataType,  DataType,    AccType, PassThrough, PassThrough, PassThrough,  ConvBwdDataFilter1x1Stride1Pad0,   2,     128,   128,   128,     4,  16,   32,   32,    4,    2,      S<4, 32, 1>,     S<1, 0, 2>,     S<1, 0, 2>,              2,             16,             16,      true,     S<4, 32, 1>,     S<2, 0, 1>,     S<0, 2, 1>,             1,              2,             16,      true,               7,               1>,
+     #if 1
+        DeviceConvndBwdDataXdl_Input_N_Di_Hi_Wi_C_Weight_K_Z_Y_X_C_Output_N_Do_Ho_Wo_K<   DataType,  DataType,  DataType,    AccType, PassThrough, PassThrough, PassThrough,  ConvBwdDataFilter1x1Stride1Pad0,   2,     256,   128,   128,     4,  16,   32,   32,    2,    2,      S<4, 64, 1>,     S<1, 0, 2>,     S<1, 0, 2>,              2,             16,             16,      true,     S<4, 64, 1>,     S<2, 0, 1>,     S<0, 2, 1>,             1,              2,             16,      true,               7,               1>,
+    #endif
+        DeviceConvndBwdDataXdl_Input_N_Di_Hi_Wi_C_Weight_K_Z_Y_X_C_Output_N_Do_Ho_Wo_K<   DataType,  DataType,  DataType,    AccType, PassThrough, PassThrough, PassThrough,  ConvBwdDataFilter1x1Stride1Pad0,   2,     128,   128,    64,     4,  16,   32,   32,    2,    2,      S<4, 32, 1>,     S<1, 0, 2>,     S<1, 0, 2>,              2,             16,             16,      true,     S<4, 32, 1>,     S<2, 0, 1>,     S<0, 2, 1>,             1,              2,             16,      true,               7,               1>,
+        DeviceConvndBwdDataXdl_Input_N_Di_Hi_Wi_C_Weight_K_Z_Y_X_C_Output_N_Do_Ho_Wo_K<   DataType,  DataType,  DataType,    AccType, PassThrough, PassThrough, PassThrough,  ConvBwdDataFilter1x1Stride1Pad0,   2,     128,    64,   128,     4,  16,   32,   32,    2,    2,      S<4, 32, 1>,     S<1, 0, 2>,     S<1, 0, 2>,              2,             16,             16,      true,     S<4, 32, 1>,     S<2, 0, 1>,     S<0, 2, 1>,             1,              4,             16,      true,               7,               1>,
+        DeviceConvndBwdDataXdl_Input_N_Di_Hi_Wi_C_Weight_K_Z_Y_X_C_Output_N_Do_Ho_Wo_K<   DataType,  DataType,  DataType,    AccType, PassThrough, PassThrough, PassThrough,  ConvBwdDataFilter1x1Stride1Pad0,   2,      64,    64,    64,     4,  16,   32,   32,    2,    2,      S<4, 16, 1>,     S<1, 0, 2>,     S<1, 0, 2>,              2,             16,             16,      true,     S<4, 16, 1>,     S<2, 0, 1>,     S<0, 2, 1>,             1,              4,             16,      true,               7,               1>,
+        DeviceConvndBwdDataXdl_Input_N_Di_Hi_Wi_C_Weight_K_Z_Y_X_C_Output_N_Do_Ho_Wo_K<   DataType,  DataType,  DataType,    AccType, PassThrough, PassThrough, PassThrough,  ConvBwdDataFilter1x1Stride1Pad0,   2,     256,   128,    64,     4,  16,   32,   32,    2,    1,      S<4, 64, 1>,     S<1, 0, 2>,     S<1, 0, 2>,              2,             16,             16,      true,     S<4, 64, 1>,     S<2, 0, 1>,     S<0, 2, 1>,             1,              1,             16,      true,               7,               1>,
+        DeviceConvndBwdDataXdl_Input_N_Di_Hi_Wi_C_Weight_K_Z_Y_X_C_Output_N_Do_Ho_Wo_K<   DataType,  DataType,  DataType,    AccType, PassThrough, PassThrough, PassThrough,  ConvBwdDataFilter1x1Stride1Pad0,   2,     256,    64,   128,     4,  16,   32,   32,    1,    2,      S<4, 64, 1>,     S<1, 0, 2>,     S<1, 0, 2>,              2,             16,             16,      true,     S<4, 64, 1>,     S<2, 0, 1>,     S<0, 2, 1>,             1,              2,             16,      true,               7,               1>,
+        DeviceConvndBwdDataXdl_Input_N_Di_Hi_Wi_C_Weight_K_Z_Y_X_C_Output_N_Do_Ho_Wo_K<   DataType,  DataType,  DataType,    AccType, PassThrough, PassThrough, PassThrough,  ConvBwdDataFilter1x1Stride1Pad0,   2,     128,   128,    32,     4,  16,   32,   32,    2,    1,      S<4, 32, 1>,     S<1, 0, 2>,     S<1, 0, 2>,              2,             16,             16,      true,     S<4, 32, 1>,     S<2, 0, 1>,     S<0, 2, 1>,             1,              1,             16,      true,               7,               1>,
+        DeviceConvndBwdDataXdl_Input_N_Di_Hi_Wi_C_Weight_K_Z_Y_X_C_Output_N_Do_Ho_Wo_K<   DataType,  DataType,  DataType,    AccType, PassThrough, PassThrough, PassThrough,  ConvBwdDataFilter1x1Stride1Pad0,   2,     128,    32,   128,     4,  16,   32,   32,    1,    2,      S<4, 32, 1>,     S<1, 0, 2>,     S<1, 0, 2>,              2,             16,             16,      true,     S<4, 32, 1>,     S<2, 0, 1>,     S<0, 2, 1>,             1,              4,             16,      true,               7,               1>,
+        DeviceConvndBwdDataXdl_Input_N_Di_Hi_Wi_C_Weight_K_Z_Y_X_C_Output_N_Do_Ho_Wo_K<   DataType,  DataType,  DataType,    AccType, PassThrough, PassThrough, PassThrough,  ConvBwdDataFilter1x1Stride1Pad0,   2,      64,    64,    32,     4,  16,   32,   32,    2,    1,      S<4, 16, 1>,     S<1, 0, 2>,     S<1, 0, 2>,              2,             16,             16,      true,     S<4, 16, 1>,     S<2, 0, 1>,     S<0, 2, 1>,             1,              2,             16,      true,               7,               1>,
+        DeviceConvndBwdDataXdl_Input_N_Di_Hi_Wi_C_Weight_K_Z_Y_X_C_Output_N_Do_Ho_Wo_K<   DataType,  DataType,  DataType,    AccType, PassThrough, PassThrough, PassThrough,  ConvBwdDataFilter1x1Stride1Pad0,   2,      64,    32,    64,     4,  16,   32,   32,    1,    2,      S<4, 16, 1>,     S<1, 0, 2>,     S<1, 0, 2>,              2,             16,             16,      true,     S<4, 16, 1>,     S<2, 0, 1>,     S<0, 2, 1>,             1,              4,             16,      true,               7,               1>
+        // clang-format on
+        >;
+
+void add_device_conv2d_bwd_data_xdl_nhwc_kyxc_nhwk_int8_instances(
+    std::vector<DeviceConvBwdDataPtr<PassThrough, PassThrough, PassThrough>>& instances)
+{
+    add_device_operation_instances(instances,
+                                   device_conv2d_bwd_data_xdl_nhwc_kyxc_nhwk_int8_instances{});
+    add_device_operation_instances(
+        instances, device_conv2d_bwd_data_xdl_nhwc_kyxc_nhwk_1x1_s1_p0_int8_instances{});
+}
+
+} // namespace device_conv2d_bwd_data_instance
+} // namespace device
+} // namespace tensor_operation
+} // namespace ck

library/src/tensor_operation_instance/gpu/convnd_bwd_data/device_conv3d_bwd_data_xdl_ndhwc_kzyxc_ndhwk_bf16_instance.cpp

+#include <stdlib.h>
+#include "config.hpp"
+#include "device_convnd_bwd_data_xdl_ndhwc_kzyxc_ndhwk.hpp"
+#include "element_wise_operation.hpp"
+#include "device_operation_instance.hpp"
+
+namespace ck {
+namespace tensor_operation {
+namespace device {
+namespace device_conv2d_bwd_data_instance {
+
+using BF16 = ushort;
+using F32  = float;
+
+template <ck::index_t... Is>
+using S = ck::Sequence<Is...>;
+
+using PassThrough = ck::tensor_operation::element_wise::PassThrough;
+static constexpr auto ConvBwdDataDefault =
+    ck::tensor_operation::device::ConvolutionBackwardDataSpecialization_t::Default;
+
+static constexpr auto ConvBwdDataFilter1x1Stride1Pad0 =
+    ck::tensor_operation::device::ConvolutionBackwardDataSpecialization_t::Filter1x1Stride1Pad0;
+
+// Compilation parameters for in[n, hi, wi, c] * wei[k, y, x, c] = out[n, ho, wo, k]
+using device_conv3d_bwd_data_xdl_ndhwc_kzyxc_ndhwk_bf16_instances =
+    std::tuple<
+        // clang-format off
+        //#############################################################################| InData| WeiData| OutData| AccData|          In|         Wei|         Out|        ConvBackward|    Num| Block|  MPer|  NPer| K0Per| K1| MPer| NPer| MXdl| NXdl|  ABlockTransfer| ABlockTransfer| ABlockTransfer| ABlockTransfer| ABlockTransfer| ABlockTransfer| ABlockLds|  BBlockTransfer| BBlockTransfer| BBlockTransfer| BlockTransfer| BBlockTransfer| BBlockTransfer| BBlockLds| CThreadTransfer| CThreadTransfer|
+        //#############################################################################|   Type|    Type|    Type|    Type| Elementwise| Elementwise| Elementwise|                Data|    Dim|  Size| Block| Block| Block|   |  XDL|  XDL|  Per|  Per|   ThreadCluster|  ThreadCluster| SrcAccessOrder|   SrcVectorDim|      SrcScalar|      DstScalar| AddExtraM|   ThreadCluster|  ThreadCluster| SrcAccessOrder|  SrcVectorDim|      SrcScalar|      DstScalar| AddExtraN| SrcDstVectorDim|       DstScalar|
+        //#############################################################################|       |        |        |        |   Operation|   Operation|   Operation|      Specialization|Spatial|      |      |      |      |   |     |     | Wave| Wave| Lengths_K0_M_K1|   ArrangeOrder|               |               |      PerVector|   PerVector_K1|          | Lengths_K0_N_K1|   ArrangeOrder|               |              |      PerVector|   PerVector_K1|          |                |       PerVector|
+        //#############################################################################|       |        |        |        |            |            |            |                    |       |      |      |      |      |   |     |     |     |     |                |               |               |               |               |               |          |                |               |               |              |               |               |          |                |                |
+        DeviceConvndBwdDataXdl_Input_N_Di_Hi_Wi_C_Weight_K_Z_Y_X_C_Output_N_Do_Ho_Wo_K<   BF16,    BF16,    BF16,     F32, PassThrough, PassThrough, PassThrough,  ConvBwdDataDefault,   3,     256,   256,   128,     4,  8,   32,   32,    4,    2,     S<4, 64, 1>,     S<1, 0, 2>,     S<1, 0, 2>,              2,              8,              8,      true,     S<4, 64, 1>,     S<2, 0, 1>,     S<0, 2, 1>,             1,              2,              8,      true,               7,               1>,
+        DeviceConvndBwdDataXdl_Input_N_Di_Hi_Wi_C_Weight_K_Z_Y_X_C_Output_N_Do_Ho_Wo_K<   BF16,    BF16,    BF16,     F32, PassThrough, PassThrough, PassThrough,  ConvBwdDataDefault,   3,     256,   128,   256,     4,  8,   32,   32,    2,    4,     S<4, 64, 1>,     S<1, 0, 2>,     S<1, 0, 2>,              2,              8,              8,      true,     S<4, 64, 1>,     S<2, 0, 1>,     S<0, 2, 1>,             1,              4,              8,      true,               7,               1>,
+        DeviceConvndBwdDataXdl_Input_N_Di_Hi_Wi_C_Weight_K_Z_Y_X_C_Output_N_Do_Ho_Wo_K<   BF16,    BF16,    BF16,     F32, PassThrough, PassThrough, PassThrough,  ConvBwdDataDefault,   3,     128,   128,   128,     4,  8,   32,   32,    4,    2,     S<4, 32, 1>,     S<1, 0, 2>,     S<1, 0, 2>,              2,              8,              8,      true,     S<4, 32, 1>,     S<2, 0, 1>,     S<0, 2, 1>,             1,              4,              8,      true,               7,               1>,
+        DeviceConvndBwdDataXdl_Input_N_Di_Hi_Wi_C_Weight_K_Z_Y_X_C_Output_N_Do_Ho_Wo_K<   BF16,    BF16,    BF16,     F32, PassThrough, PassThrough, PassThrough,  ConvBwdDataDefault,   3,     256,   128,   128,     4,  8,   32,   32,    2,    2,     S<4, 64, 1>,     S<1, 0, 2>,     S<1, 0, 2>,              2,              8,              8,      true,     S<4, 64, 1>,     S<2, 0, 1>,     S<0, 2, 1>,             1,              2,              8,      true,               7,               1>,
+        DeviceConvndBwdDataXdl_Input_N_Di_Hi_Wi_C_Weight_K_Z_Y_X_C_Output_N_Do_Ho_Wo_K<   BF16,    BF16,    BF16,     F32, PassThrough, PassThrough, PassThrough,  ConvBwdDataDefault,   3,     128,   128,    64,     4,  8,   32,   32,    2,    2,     S<4, 32, 1>,     S<1, 0, 2>,     S<1, 0, 2>,              2,              8,              8,      true,     S<4, 32, 1>,     S<2, 0, 1>,     S<0, 2, 1>,             1,              2,              8,      true,               7,               1>,
+        DeviceConvndBwdDataXdl_Input_N_Di_Hi_Wi_C_Weight_K_Z_Y_X_C_Output_N_Do_Ho_Wo_K<   BF16,    BF16,    BF16,     F32, PassThrough, PassThrough, PassThrough,  ConvBwdDataDefault,   3,     128,    64,   128,     4,  8,   32,   32,    2,    2,     S<4, 32, 1>,     S<1, 0, 2>,     S<1, 0, 2>,              2,              8,              8,      true,     S<4, 32, 1>,     S<2, 0, 1>,     S<0, 2, 1>,             1,              4,              8,      true,               7,               1>,
+        DeviceConvndBwdDataXdl_Input_N_Di_Hi_Wi_C_Weight_K_Z_Y_X_C_Output_N_Do_Ho_Wo_K<   BF16,    BF16,    BF16,     F32, PassThrough, PassThrough, PassThrough,  ConvBwdDataDefault,   3,      64,    64,    64,     4,  8,   32,   32,    2,    2,     S<4, 16, 1>,     S<1, 0, 2>,     S<1, 0, 2>,              2,              8,              8,      true,     S<4, 16, 1>,     S<2, 0, 1>,     S<0, 2, 1>,             1,              4,              8,      true,               7,               1>,
+        DeviceConvndBwdDataXdl_Input_N_Di_Hi_Wi_C_Weight_K_Z_Y_X_C_Output_N_Do_Ho_Wo_K<   BF16,    BF16,    BF16,     F32, PassThrough, PassThrough, PassThrough,  ConvBwdDataDefault,   3,     256,   128,    64,     4,  8,   32,   32,    2,    1,     S<4, 64, 1>,     S<1, 0, 2>,     S<1, 0, 2>,              2,              8,              8,      true,     S<4, 64, 1>,     S<2, 0, 1>,     S<0, 2, 1>,             1,              1,              8,      true,               7,               1>,
+        DeviceConvndBwdDataXdl_Input_N_Di_Hi_Wi_C_Weight_K_Z_Y_X_C_Output_N_Do_Ho_Wo_K<   BF16,    BF16,    BF16,     F32, PassThrough, PassThrough, PassThrough,  ConvBwdDataDefault,   3,     256,    64,   128,     4,  8,   32,   32,    1,    2,     S<4, 64, 1>,     S<1, 0, 2>,     S<1, 0, 2>,              2,              8,              8,      true,     S<4, 64, 1>,     S<2, 0, 1>,     S<0, 2, 1>,             1,              2,              8,      true,               7,               1>,
+        DeviceConvndBwdDataXdl_Input_N_Di_Hi_Wi_C_Weight_K_Z_Y_X_C_Output_N_Do_Ho_Wo_K<   BF16,    BF16,    BF16,     F32, PassThrough, PassThrough, PassThrough,  ConvBwdDataDefault,   3,     128,   128,    32,     4,  8,   32,   32,    2,    1,     S<4, 32, 1>,     S<1, 0, 2>,     S<1, 0, 2>,              2,              8,              8,      true,     S<4, 32, 1>,     S<2, 0, 1>,     S<0, 2, 1>,             1,              1,              8,      true,               7,               1>,
+        DeviceConvndBwdDataXdl_Input_N_Di_Hi_Wi_C_Weight_K_Z_Y_X_C_Output_N_Do_Ho_Wo_K<   BF16,    BF16,    BF16,     F32, PassThrough, PassThrough, PassThrough,  ConvBwdDataDefault,   3,     128,    32,   128,     4,  8,   32,   32,    1,    2,     S<4, 32, 1>,     S<1, 0, 2>,     S<1, 0, 2>,              2,              8,              8,      true,     S<4, 32, 1>,     S<2, 0, 1>,     S<0, 2, 1>,             1,              4,              8,      true,               7,               1>,
+        DeviceConvndBwdDataXdl_Input_N_Di_Hi_Wi_C_Weight_K_Z_Y_X_C_Output_N_Do_Ho_Wo_K<   BF16,    BF16,    BF16,     F32, PassThrough, PassThrough, PassThrough,  ConvBwdDataDefault,   3,      64,    64,    32,     4,  8,   32,   32,    2,    1,     S<4, 16, 1>,     S<1, 0, 2>,     S<1, 0, 2>,              2,              8,              8,      true,     S<4, 16, 1>,     S<2, 0, 1>,     S<0, 2, 1>,             1,              2,              8,      true,               7,               1>,
+        DeviceConvndBwdDataXdl_Input_N_Di_Hi_Wi_C_Weight_K_Z_Y_X_C_Output_N_Do_Ho_Wo_K<   BF16,    BF16,    BF16,     F32, PassThrough, PassThrough, PassThrough,  ConvBwdDataDefault,   3,      64,    32,    64,     4,  8,   32,   32,    1,    2,     S<4, 16, 1>,     S<1, 0, 2>,     S<1, 0, 2>,              2,              8,              8,      true,     S<4, 16, 1>,     S<2, 0, 1>,     S<0, 2, 1>,             1,              2,              8,      true,               7,               1>
+        // clang-format on
+        >;
+
+using device_conv3d_bwd_data_xdl_ndhwc_kzyxc_ndhwk_1x1_s1_p0_bf16_instances =
+    std::tuple<
+        // clang-format off
+        //#############################################################################| InData| WeiData| OutData| AccData|          In|         Wei|         Out|                     ConvBackward|    Num| Block|  MPer|  NPer| K0Per| K1| MPer| NPer| MXdl| NXdl|  ABlockTransfer| ABlockTransfer| ABlockTransfer| ABlockTransfer| ABlockTransfer| ABlockTransfer| ABlockLds|  BBlockTransfer| BBlockTransfer| BBlockTransfer| BlockTransfer| BBlockTransfer| BBlockTransfer| BBlockLds| CThreadTransfer| CThreadTransfer|
+        //#############################################################################|   Type|    Type|    Type|    Type| Elementwise| Elementwise| Elementwise|                             Data|    Dim|  Size| Block| Block| Block|   |  XDL|  XDL|  Per|  Per|   ThreadCluster|  ThreadCluster| SrcAccessOrder|   SrcVectorDim|      SrcScalar|      DstScalar| AddExtraM|   ThreadCluster|  ThreadCluster| SrcAccessOrder|  SrcVectorDim|      SrcScalar|      DstScalar| AddExtraN| SrcDstVectorDim|       DstScalar|
+        //#############################################################################|       |        |        |        |   Operation|   Operation|   Operation|                   Specialization|Spatial|      |      |      |      |   |     |     | Wave| Wave| Lengths_K0_M_K1|   ArrangeOrder|               |        ./       |      PerVector|   PerVector_K1|          | Lengths_K0_N_K1|   ArrangeOrder|               |              |      PerVector|   PerVector_K1|          |                |       PerVector|
+        //#############################################################################|       |        |        |        |            |            |            |                                 |       |      |      |      |      |   |     |     |     |     |                |               |               |               |               |               |          |                |               |               |              |               |               |          |                |                |
+        DeviceConvndBwdDataXdl_Input_N_Di_Hi_Wi_C_Weight_K_Z_Y_X_C_Output_N_Do_Ho_Wo_K<   BF16,    BF16,    BF16,     F32, PassThrough, PassThrough, PassThrough,  ConvBwdDataFilter1x1Stride1Pad0,   3,     256,   256,   128,     4,  8,   32,   32,    4,    2,     S<4, 64, 1>,     S<1, 0, 2>,     S<1, 0, 2>,              2,              8,              8,      true,     S<4, 64, 1>,     S<2, 0, 1>,     S<0, 2, 1>,             1,              2,              8,      true,               7,               1>,
+        DeviceConvndBwdDataXdl_Input_N_Di_Hi_Wi_C_Weight_K_Z_Y_X_C_Output_N_Do_Ho_Wo_K<   BF16,    BF16,    BF16,     F32, PassThrough, PassThrough, PassThrough,  ConvBwdDataFilter1x1Stride1Pad0,   3,     256,   128,   256,     4,  8,   32,   32,    2,    4,     S<4, 64, 1>,     S<1, 0, 2>,     S<1, 0, 2>,              2,              8,              8,      true,     S<4, 64, 1>,     S<2, 0, 1>,     S<0, 2, 1>,             1,              4,              8,      true,               7,               1>,
+        DeviceConvndBwdDataXdl_Input_N_Di_Hi_Wi_C_Weight_K_Z_Y_X_C_Output_N_Do_Ho_Wo_K<   BF16,    BF16,    BF16,     F32, PassThrough, PassThrough, PassThrough,  ConvBwdDataFilter1x1Stride1Pad0,   3,     128,   128,   128,     4,  8,   32,   32,    4,    2,     S<4, 32, 1>,     S<1, 0, 2>,     S<1, 0, 2>,              2,              8,              8,      true,     S<4, 32, 1>,     S<2, 0, 1>,     S<0, 2, 1>,             1,              4,              8,      true,               7,               1>,
+        DeviceConvndBwdDataXdl_Input_N_Di_Hi_Wi_C_Weight_K_Z_Y_X_C_Output_N_Do_Ho_Wo_K<   BF16,    BF16,    BF16,     F32, PassThrough, PassThrough, PassThrough,  ConvBwdDataFilter1x1Stride1Pad0,   3,     256,   128,   128,     4,  8,   32,   32,    2,    2,     S<4, 64, 1>,     S<1, 0, 2>,     S<1, 0, 2>,              2,              8,              8,      true,     S<4, 64, 1>,     S<2, 0, 1>,     S<0, 2, 1>,             1,              2,              8,      true,               7,               1>,
+        DeviceConvndBwdDataXdl_Input_N_Di_Hi_Wi_C_Weight_K_Z_Y_X_C_Output_N_Do_Ho_Wo_K<   BF16,    BF16,    BF16,     F32, PassThrough, PassThrough, PassThrough,  ConvBwdDataFilter1x1Stride1Pad0,   3,     128,   128,    64,     4,  8,   32,   32,    2,    2,     S<4, 32, 1>,     S<1, 0, 2>,     S<1, 0, 2>,              2,              8,              8,      true,     S<4, 32, 1>,     S<2, 0, 1>,     S<0, 2, 1>,             1,              2,              8,      true,               7,               1>,
+        DeviceConvndBwdDataXdl_Input_N_Di_Hi_Wi_C_Weight_K_Z_Y_X_C_Output_N_Do_Ho_Wo_K<   BF16,    BF16,    BF16,     F32, PassThrough, PassThrough, PassThrough,  ConvBwdDataFilter1x1Stride1Pad0,   3,     128,    64,   128,     4,  8,   32,   32,    2,    2,     S<4, 32, 1>,     S<1, 0, 2>,     S<1, 0, 2>,              2,              8,              8,      true,     S<4, 32, 1>,     S<2, 0, 1>,     S<0, 2, 1>,             1,              4,              8,      true,               7,               1>,
+        DeviceConvndBwdDataXdl_Input_N_Di_Hi_Wi_C_Weight_K_Z_Y_X_C_Output_N_Do_Ho_Wo_K<   BF16,    BF16,    BF16,     F32, PassThrough, PassThrough, PassThrough,  ConvBwdDataFilter1x1Stride1Pad0,   3,      64,    64,    64,     4,  8,   32,   32,    2,    2,     S<4, 16, 1>,     S<1, 0, 2>,     S<1, 0, 2>,              2,              8,              8,      true,     S<4, 16, 1>,     S<2, 0, 1>,     S<0, 2, 1>,             1,              4,              8,      true,               7,               1>,
+        DeviceConvndBwdDataXdl_Input_N_Di_Hi_Wi_C_Weight_K_Z_Y_X_C_Output_N_Do_Ho_Wo_K<   BF16,    BF16,    BF16,     F32, PassThrough, PassThrough, PassThrough,  ConvBwdDataFilter1x1Stride1Pad0,   3,     256,   128,    64,     4,  8,   32,   32,    2,    1,     S<4, 64, 1>,     S<1, 0, 2>,     S<1, 0, 2>,              2,              8,              8,      true,     S<4, 64, 1>,     S<2, 0, 1>,     S<0, 2, 1>,             1,              1,              8,      true,               7,               1>,
+        DeviceConvndBwdDataXdl_Input_N_Di_Hi_Wi_C_Weight_K_Z_Y_X_C_Output_N_Do_Ho_Wo_K<   BF16,    BF16,    BF16,     F32, PassThrough, PassThrough, PassThrough,  ConvBwdDataFilter1x1Stride1Pad0,   3,     256,    64,   128,     4,  8,   32,   32,    1,    2,     S<4, 64, 1>,     S<1, 0, 2>,     S<1, 0, 2>,              2,              8,              8,      true,     S<4, 64, 1>,     S<2, 0, 1>,     S<0, 2, 1>,             1,              2,              8,      true,               7,               1>,
+        DeviceConvndBwdDataXdl_Input_N_Di_Hi_Wi_C_Weight_K_Z_Y_X_C_Output_N_Do_Ho_Wo_K<   BF16,    BF16,    BF16,     F32, PassThrough, PassThrough, PassThrough,  ConvBwdDataFilter1x1Stride1Pad0,   3,     128,   128,    32,     4,  8,   32,   32,    2,    1,     S<4, 32, 1>,     S<1, 0, 2>,     S<1, 0, 2>,              2,              8,              8,      true,     S<4, 32, 1>,     S<2, 0, 1>,     S<0, 2, 1>,             1,              1,              8,      true,               7,               1>,
+        DeviceConvndBwdDataXdl_Input_N_Di_Hi_Wi_C_Weight_K_Z_Y_X_C_Output_N_Do_Ho_Wo_K<   BF16,    BF16,    BF16,     F32, PassThrough, PassThrough, PassThrough,  ConvBwdDataFilter1x1Stride1Pad0,   3,     128,    32,   128,     4,  8,   32,   32,    1,    2,     S<4, 32, 1>,     S<1, 0, 2>,     S<1, 0, 2>,              2,              8,              8,      true,     S<4, 32, 1>,     S<2, 0, 1>,     S<0, 2, 1>,             1,              4,              8,      true,               7,               1>,
+        DeviceConvndBwdDataXdl_Input_N_Di_Hi_Wi_C_Weight_K_Z_Y_X_C_Output_N_Do_Ho_Wo_K<   BF16,    BF16,    BF16,     F32, PassThrough, PassThrough, PassThrough,  ConvBwdDataFilter1x1Stride1Pad0,   3,      64,    64,    32,     4,  8,   32,   32,    2,    1,     S<4, 16, 1>,     S<1, 0, 2>,     S<1, 0, 2>,              2,              8,              8,      true,     S<4, 16, 1>,     S<2, 0, 1>,     S<0, 2, 1>,             1,              2,              8,      true,               7,               1>,
+        DeviceConvndBwdDataXdl_Input_N_Di_Hi_Wi_C_Weight_K_Z_Y_X_C_Output_N_Do_Ho_Wo_K<   BF16,    BF16,    BF16,     F32, PassThrough, PassThrough, PassThrough,  ConvBwdDataFilter1x1Stride1Pad0,   3,      64,    32,    64,     4,  8,   32,   32,    1,    2,     S<4, 16, 1>,     S<1, 0, 2>,     S<1, 0, 2>,              2,              8,              8,      true,     S<4, 16, 1>,     S<2, 0, 1>,     S<0, 2, 1>,             1,              2,              8,      true,               7,               1>
+        // clang-format on
+        >;
+
+void add_device_conv3d_bwd_data_xdl_ndhwc_kzyxc_ndhwk_bf16_instances(
+    std::vector<DeviceConvBwdDataPtr<PassThrough, PassThrough, PassThrough>>& instances)
+{
+    add_device_operation_instances(instances,
+                                   device_conv3d_bwd_data_xdl_ndhwc_kzyxc_ndhwk_bf16_instances{});
+    add_device_operation_instances(
+        instances, device_conv3d_bwd_data_xdl_ndhwc_kzyxc_ndhwk_1x1_s1_p0_bf16_instances{});
+}
+
+} // namespace device_conv2d_bwd_data_instance
+} // namespace device
+} // namespace tensor_operation
+} // namespace ck

library/src/tensor_operation_instance/gpu/convnd_bwd_data/device_conv3d_bwd_data_xdl_ndhwc_kzyxc_ndhwk_f16_instance.cpp

+#include <stdlib.h>
+#include "config.hpp"
+#include "device_convnd_bwd_data_xdl_ndhwc_kzyxc_ndhwk.hpp"
+#include "element_wise_operation.hpp"
+#include "device_operation_instance.hpp"
+
+namespace ck {
+namespace tensor_operation {
+namespace device {
+namespace device_conv2d_bwd_data_instance {
+
+using F16 = ck::half_t;
+using F32 = float;
+
+template <ck::index_t... Is>
+using S = ck::Sequence<Is...>;
+
+using PassThrough = ck::tensor_operation::element_wise::PassThrough;
+static constexpr auto ConvBwdDataDefault =
+    ck::tensor_operation::device::ConvolutionBackwardDataSpecialization_t::Default;
+
+static constexpr auto ConvBwdDataFilter1x1Stride1Pad0 =
+    ck::tensor_operation::device::ConvolutionBackwardDataSpecialization_t::Filter1x1Stride1Pad0;
+
+// Compilation parameters for in[n, hi, wi, c] * wei[k, y, x, c] = out[n, ho, wo, k]
+using device_conv3d_bwd_data_xdl_ndhwc_kzyxc_ndhwk_f16_instances =
+    std::tuple<
+        // clang-format off
+        //#############################################################################| InData| WeiData| OutData| AccData|          In|         Wei|         Out|       ConvBackward|    Num| Block|  MPer|  NPer| K0Per| K1| MPer| NPer| MXdl| NXdl|  ABlockTransfer| ABlockTransfer| ABlockTransfer| ABlockTransfer| ABlockTransfer| ABlockTransfer| ABlockLds|  BBlockTransfer| BBlockTransfer| BBlockTransfer| BlockTransfer| BBlockTransfer| BBlockTransfer| BBlockLds| CThreadTransfer| CThreadTransfer|
+        //#############################################################################|   Type|    Type|    Type|    Type| Elementwise| Elementwise| Elementwise|               Data|    Dim|  Size| Block| Block| Block|   |  XDL|  XDL|  Per|  Per|   ThreadCluster|  ThreadCluster| SrcAccessOrder|   SrcVectorDim|      SrcScalar|      DstScalar| AddExtraM|   ThreadCluster|  ThreadCluster| SrcAccessOrder|  SrcVectorDim|      SrcScalar|      DstScalar| AddExtraN| SrcDstVectorDim|       DstScalar|
+        //#############################################################################|       |        |        |        |   Operation|   Operation|   Operation|     Specialization|Spatial|      |      |      |      |   |     |     | Wave| Wave| Lengths_K0_M_K1|   ArrangeOrder|               |               |      PerVector|   PerVector_K1|          | Lengths_K0_N_K1|   ArrangeOrder|               |              |      PerVector|   PerVector_K1|          |                |       PerVector|
+        //#############################################################################|       |        |        |        |            |            |            |                   |       |      |      |      |      |   |     |     |     |     |                |               |               |               |               |               |          |                |               |               |              |               |               |          |                |                |
+        DeviceConvndBwdDataXdl_Input_N_Di_Hi_Wi_C_Weight_K_Z_Y_X_C_Output_N_Do_Ho_Wo_K<    F16,     F16,     F16,     F32, PassThrough, PassThrough, PassThrough, ConvBwdDataDefault,   3,     256,   256,   128,     4,   8,  32,   32,    4,    2,     S<4, 64, 1>,     S<1, 0, 2>,     S<1, 0, 2>,              2,              8,              8,      true,     S<4, 64, 1>,     S<2, 0, 1>,     S<0, 2, 1>,             1,              2,              8,      true,               7,               1>,
+        DeviceConvndBwdDataXdl_Input_N_Di_Hi_Wi_C_Weight_K_Z_Y_X_C_Output_N_Do_Ho_Wo_K<    F16,     F16,     F16,     F32, PassThrough, PassThrough, PassThrough, ConvBwdDataDefault,   3,     128,   128,   128,     4,   8,  32,   32,    4,    2,     S<4, 32, 1>,     S<1, 0, 2>,     S<1, 0, 2>,              2,              8,              8,      true,     S<4, 32, 1>,     S<2, 0, 1>,     S<0, 2, 1>,             1,              4,              8,      true,               7,               1>,
+#if 1
+        DeviceConvndBwdDataXdl_Input_N_Di_Hi_Wi_C_Weight_K_Z_Y_X_C_Output_N_Do_Ho_Wo_K<    F16,     F16,     F16,     F32, PassThrough, PassThrough, PassThrough, ConvBwdDataDefault,   3,     256,   128,   256,     4,   8,  32,   32,    2,    4,     S<4, 64, 1>,     S<1, 0, 2>,     S<1, 0, 2>,              2,              8,              8,      true,     S<4, 64, 1>,     S<2, 0, 1>,     S<0, 2, 1>,             1,              4,              8,      true,               7,               1>,  
+        DeviceConvndBwdDataXdl_Input_N_Di_Hi_Wi_C_Weight_K_Z_Y_X_C_Output_N_Do_Ho_Wo_K<    F16,     F16,     F16,     F32, PassThrough, PassThrough, PassThrough, ConvBwdDataDefault,   3,     256,   128,   128,     4,   8,  32,   32,    2,    2,     S<4, 64, 1>,     S<1, 0, 2>,     S<1, 0, 2>,              2,              8,              8,      true,     S<4, 64, 1>,     S<2, 0, 1>,     S<0, 2, 1>,             1,              2,              8,      true,               7,               1>,
+        DeviceConvndBwdDataXdl_Input_N_Di_Hi_Wi_C_Weight_K_Z_Y_X_C_Output_N_Do_Ho_Wo_K<    F16,     F16,     F16,     F32, PassThrough, PassThrough, PassThrough, ConvBwdDataDefault,   3,     256,    64,   128,     4,   8,  32,   32,    1,    2,     S<4, 64, 1>,     S<1, 0, 2>,     S<1, 0, 2>,              2,              8,              8,      true,     S<4, 64, 1>,     S<2, 0, 1>,     S<0, 2, 1>,             1,              2,              8,      true,               7,               1>,
+        DeviceConvndBwdDataXdl_Input_N_Di_Hi_Wi_C_Weight_K_Z_Y_X_C_Output_N_Do_Ho_Wo_K<    F16,     F16,     F16,     F32, PassThrough, PassThrough, PassThrough, ConvBwdDataDefault,   3,     128,    32,   128,     4,   8,  32,   32,    1,    2,     S<4, 32, 1>,     S<1, 0, 2>,     S<1, 0, 2>,              2,              8,              8,      true,     S<4, 32, 1>,     S<2, 0, 1>,     S<0, 2, 1>,             1,              4,              8,      true,               7,               1>,
+        DeviceConvndBwdDataXdl_Input_N_Di_Hi_Wi_C_Weight_K_Z_Y_X_C_Output_N_Do_Ho_Wo_K<    F16,     F16,     F16,     F32, PassThrough, PassThrough, PassThrough, ConvBwdDataDefault,   3,     128,    64,   128,     4,   8,  32,   32,    2,    2,     S<4, 32, 1>,     S<1, 0, 2>,     S<1, 0, 2>,              2,              8,              8,      true,     S<4, 32, 1>,     S<2, 0, 1>,     S<0, 2, 1>,             1,              4,              8,      true,               7,               1>,
+        DeviceConvndBwdDataXdl_Input_N_Di_Hi_Wi_C_Weight_K_Z_Y_X_C_Output_N_Do_Ho_Wo_K<    F16,     F16,     F16,     F32, PassThrough, PassThrough, PassThrough, ConvBwdDataDefault,   3,     256,   128,    64,     4,   8,  32,   32,    2,    1,     S<4, 64, 1>,     S<1, 0, 2>,     S<1, 0, 2>,              2,              8,              8,      true,     S<4, 64, 1>,     S<2, 0, 1>,     S<0, 2, 1>,             1,              1,              8,      true,               7,               1>,
+        DeviceConvndBwdDataXdl_Input_N_Di_Hi_Wi_C_Weight_K_Z_Y_X_C_Output_N_Do_Ho_Wo_K<    F16,     F16,     F16,     F32, PassThrough, PassThrough, PassThrough, ConvBwdDataDefault,   3,      64,    32,    64,     4,   8,  32,   32,    1,    2,     S<4, 16, 1>,     S<1, 0, 2>,     S<1, 0, 2>,              2,              8,              8,      true,     S<4, 16, 1>,     S<2, 0, 1>,     S<0, 2, 1>,             1,              2,              8,      true,               7,               1>,
+#endif
+        DeviceConvndBwdDataXdl_Input_N_Di_Hi_Wi_C_Weight_K_Z_Y_X_C_Output_N_Do_Ho_Wo_K<    F16,     F16,     F16,     F32, PassThrough, PassThrough, PassThrough, ConvBwdDataDefault,   3,     128,   128,    64,     4,   8,  32,   32,    2,    2,     S<4, 32, 1>,     S<1, 0, 2>,     S<1, 0, 2>,              2,              8,              8,      true,     S<4, 32, 1>,     S<2, 0, 1>,     S<0, 2, 1>,             1,              2,              8,      true,               7,               1>,
+        DeviceConvndBwdDataXdl_Input_N_Di_Hi_Wi_C_Weight_K_Z_Y_X_C_Output_N_Do_Ho_Wo_K<    F16,     F16,     F16,     F32, PassThrough, PassThrough, PassThrough, ConvBwdDataDefault,   3,      64,    64,    64,     4,   8,  32,   32,    2,    2,     S<4, 16, 1>,     S<1, 0, 2>,     S<1, 0, 2>,              2,              8,              8,      true,     S<4, 16, 1>,     S<2, 0, 1>,     S<0, 2, 1>,             1,              4,              8,      true,               7,               1>,
+        DeviceConvndBwdDataXdl_Input_N_Di_Hi_Wi_C_Weight_K_Z_Y_X_C_Output_N_Do_Ho_Wo_K<    F16,     F16,     F16,     F32, PassThrough, PassThrough, PassThrough, ConvBwdDataDefault,   3,     128,   128,    32,     4,   8,  32,   32,    2,    1,     S<4, 32, 1>,     S<1, 0, 2>,     S<1, 0, 2>,              2,              8,              8,      true,     S<4, 32, 1>,     S<2, 0, 1>,     S<0, 2, 1>,             1,              1,              8,      true,               7,               1>,
+        DeviceConvndBwdDataXdl_Input_N_Di_Hi_Wi_C_Weight_K_Z_Y_X_C_Output_N_Do_Ho_Wo_K<    F16,     F16,     F16,     F32, PassThrough, PassThrough, PassThrough, ConvBwdDataDefault,   3,      64,    64,    32,     4,   8,  32,   32,    2,    1,     S<4, 16, 1>,     S<1, 0, 2>,     S<1, 0, 2>,              2,              8,              8,      true,     S<4, 16, 1>,     S<2, 0, 1>,     S<0, 2, 1>,             1,              2,              8,      true,               7,               1>
+        // clang-format on
+        >;
+
+using device_conv3d_bwd_data_xdl_ndhwc_kzyxc_ndhwk_1x1_s1_p0_f16_instances =
+    std::tuple<
+        // clang-format off
+        //#############################################################################| InData| WeiData| OutData| AccData|          In|         Wei|         Out|                     ConvBackward|    Num| Block|  MPer|  NPer| K0Per| K1| MPer| NPer| MXdl| NXdl|  ABlockTransfer| ABlockTransfer| ABlockTransfer| ABlockTransfer| ABlockTransfer| ABlockTransfer| ABlockLds|  BBlockTransfer| BBlockTransfer| BBlockTransfer| BlockTransfer| BBlockTransfer| BBlockTransfer| BBlockLds| CThreadTransfer| CThreadTransfer|
+        //#############################################################################|   Type|    Type|    Type|    Type| Elementwise| Elementwise| Elementwise|                             Data|    Dim|  Size| Block| Block| Block|   |  XDL|  XDL|  Per|  Per|   ThreadCluster|  ThreadCluster| SrcAccessOrder|   SrcVectorDim|      SrcScalar|      DstScalar| AddExtraM|   ThreadCluster|  ThreadCluster| SrcAccessOrder|  SrcVectorDim|      SrcScalar|      DstScalar| AddExtraN| SrcDstVectorDim|       DstScalar|
+        //#############################################################################|       |        |        |        |   Operation|   Operation|   Operation|                   Specialization|Spatial|      |      |      |      |   |     |     | Wave| Wave| Lengths_K0_M_K1|   ArrangeOrder|               |               |      PerVector|   PerVector_K1|          | Lengths_K0_N_K1|   ArrangeOrder|               |              |      PerVector|   PerVector_K1|          |                |       PerVector|
+        //#############################################################################|       |        |        |        |            |            |            |                                 |       |      |      |      |      |   |     |     |     |     |                |               |               |               |               |               |          |                |               |               |              |               |               |          |                |                |
+        DeviceConvndBwdDataXdl_Input_N_Di_Hi_Wi_C_Weight_K_Z_Y_X_C_Output_N_Do_Ho_Wo_K<    F16,     F16,     F16,     F32, PassThrough, PassThrough, PassThrough,  ConvBwdDataFilter1x1Stride1Pad0,   3,     256,   256,   128,     4,  8,   32,   32,    4,    2,     S<4, 64, 1>,     S<1, 0, 2>,     S<1, 0, 2>,              2,              8,              8,      true,     S<4, 64, 1>,     S<2, 0, 1>,     S<0, 2, 1>,             1,              2,              8,      true,               7,               1>,
+        DeviceConvndBwdDataXdl_Input_N_Di_Hi_Wi_C_Weight_K_Z_Y_X_C_Output_N_Do_Ho_Wo_K<    F16,     F16,     F16,     F32, PassThrough, PassThrough, PassThrough,  ConvBwdDataFilter1x1Stride1Pad0,   3,     256,   128,   256,     4,  8,   32,   32,    2,    4,     S<4, 64, 1>,     S<1, 0, 2>,     S<1, 0, 2>,              2,              8,              8,      true,     S<4, 64, 1>,     S<2, 0, 1>,     S<0, 2, 1>,             1,              4,              8,      true,               7,               1>,
+        DeviceConvndBwdDataXdl_Input_N_Di_Hi_Wi_C_Weight_K_Z_Y_X_C_Output_N_Do_Ho_Wo_K<    F16,     F16,     F16,     F32, PassThrough, PassThrough, PassThrough,  ConvBwdDataFilter1x1Stride1Pad0,   3,     128,   128,   128,     4,  8,   32,   32,    4,    2,     S<4, 32, 1>,     S<1, 0, 2>,     S<1, 0, 2>,              2,              8,              8,      true,     S<4, 32, 1>,     S<2, 0, 1>,     S<0, 2, 1>,             1,              4,              8,      true,               7,               1>,
+        DeviceConvndBwdDataXdl_Input_N_Di_Hi_Wi_C_Weight_K_Z_Y_X_C_Output_N_Do_Ho_Wo_K<    F16,     F16,     F16,     F32, PassThrough, PassThrough, PassThrough,  ConvBwdDataFilter1x1Stride1Pad0,   3,     256,   128,   128,     4,  8,   32,   32,    2,    2,     S<4, 64, 1>,     S<1, 0, 2>,     S<1, 0, 2>,              2,              8,              8,      true,     S<4, 64, 1>,     S<2, 0, 1>,     S<0, 2, 1>,             1,              2,              8,      true,               7,               1>,
+        DeviceConvndBwdDataXdl_Input_N_Di_Hi_Wi_C_Weight_K_Z_Y_X_C_Output_N_Do_Ho_Wo_K<    F16,     F16,     F16,     F32, PassThrough, PassThrough, PassThrough,  ConvBwdDataFilter1x1Stride1Pad0,   3,     128,   128,    64,     4,  8,   32,   32,    2,    2,     S<4, 32, 1>,     S<1, 0, 2>,     S<1, 0, 2>,              2,              8,              8,      true,     S<4, 32, 1>,     S<2, 0, 1>,     S<0, 2, 1>,             1,              2,              8,      true,               7,               1>,
+        DeviceConvndBwdDataXdl_Input_N_Di_Hi_Wi_C_Weight_K_Z_Y_X_C_Output_N_Do_Ho_Wo_K<    F16,     F16,     F16,     F32, PassThrough, PassThrough, PassThrough,  ConvBwdDataFilter1x1Stride1Pad0,   3,     128,    64,   128,     4,  8,   32,   32,    2,    2,     S<4, 32, 1>,     S<1, 0, 2>,     S<1, 0, 2>,              2,              8,              8,      true,     S<4, 32, 1>,     S<2, 0, 1>,     S<0, 2, 1>,             1,              4,              8,      true,               7,               1>,
+        DeviceConvndBwdDataXdl_Input_N_Di_Hi_Wi_C_Weight_K_Z_Y_X_C_Output_N_Do_Ho_Wo_K<    F16,     F16,     F16,     F32, PassThrough, PassThrough, PassThrough,  ConvBwdDataFilter1x1Stride1Pad0,   3,      64,    64,    64,     4,  8,   32,   32,    2,    2,     S<4, 16, 1>,     S<1, 0, 2>,     S<1, 0, 2>,              2,              8,              8,      true,     S<4, 16, 1>,     S<2, 0, 1>,     S<0, 2, 1>,             1,              4,              8,      true,               7,               1>,
+        DeviceConvndBwdDataXdl_Input_N_Di_Hi_Wi_C_Weight_K_Z_Y_X_C_Output_N_Do_Ho_Wo_K<    F16,     F16,     F16,     F32, PassThrough, PassThrough, PassThrough,  ConvBwdDataFilter1x1Stride1Pad0,   3,     256,   128,    64,     4,  8,   32,   32,    2,    1,     S<4, 64, 1>,     S<1, 0, 2>,     S<1, 0, 2>,              2,              8,              8,      true,     S<4, 64, 1>,     S<2, 0, 1>,     S<0, 2, 1>,             1,              1,              8,      true,               7,               1>,
+        DeviceConvndBwdDataXdl_Input_N_Di_Hi_Wi_C_Weight_K_Z_Y_X_C_Output_N_Do_Ho_Wo_K<    F16,     F16,     F16,     F32, PassThrough, PassThrough, PassThrough,  ConvBwdDataFilter1x1Stride1Pad0,   3,     256,    64,   128,     4,  8,   32,   32,    1,    2,     S<4, 64, 1>,     S<1, 0, 2>,     S<1, 0, 2>,              2,              8,              8,      true,     S<4, 64, 1>,     S<2, 0, 1>,     S<0, 2, 1>,             1,              2,              8,      true,               7,               1>,
+        DeviceConvndBwdDataXdl_Input_N_Di_Hi_Wi_C_Weight_K_Z_Y_X_C_Output_N_Do_Ho_Wo_K<    F16,     F16,     F16,     F32, PassThrough, PassThrough, PassThrough,  ConvBwdDataFilter1x1Stride1Pad0,   3,     128,   128,    32,     4,  8,   32,   32,    2,    1,     S<4, 32, 1>,     S<1, 0, 2>,     S<1, 0, 2>,              2,              8,              8,      true,     S<4, 32, 1>,     S<2, 0, 1>,     S<0, 2, 1>,             1,              1,              8,      true,               7,               1>,
+        DeviceConvndBwdDataXdl_Input_N_Di_Hi_Wi_C_Weight_K_Z_Y_X_C_Output_N_Do_Ho_Wo_K<    F16,     F16,     F16,     F32, PassThrough, PassThrough, PassThrough,  ConvBwdDataFilter1x1Stride1Pad0,   3,     128,    32,   128,     4,  8,   32,   32,    1,    2,     S<4, 32, 1>,     S<1, 0, 2>,     S<1, 0, 2>,              2,              8,              8,      true,     S<4, 32, 1>,     S<2, 0, 1>,     S<0, 2, 1>,             1,              4,              8,      true,               7,               1>,
+        DeviceConvndBwdDataXdl_Input_N_Di_Hi_Wi_C_Weight_K_Z_Y_X_C_Output_N_Do_Ho_Wo_K<    F16,     F16,     F16,     F32, PassThrough, PassThrough, PassThrough,  ConvBwdDataFilter1x1Stride1Pad0,   3,      64,    64,    32,     4,  8,   32,   32,    2,    1,     S<4, 16, 1>,     S<1, 0, 2>,     S<1, 0, 2>,              2,              8,              8,      true,     S<4, 16, 1>,     S<2, 0, 1>,     S<0, 2, 1>,             1,              2,              8,      true,               7,               1>,
+        DeviceConvndBwdDataXdl_Input_N_Di_Hi_Wi_C_Weight_K_Z_Y_X_C_Output_N_Do_Ho_Wo_K<    F16,     F16,     F16,     F32, PassThrough, PassThrough, PassThrough,  ConvBwdDataFilter1x1Stride1Pad0,   3,      64,    32,    64,     4,  8,   32,   32,    1,    2,     S<4, 16, 1>,     S<1, 0, 2>,     S<1, 0, 2>,              2,              8,              8,      true,     S<4, 16, 1>,     S<2, 0, 1>,     S<0, 2, 1>,             1,              2,              8,      true,               7,               1>
+        // clang-format on
+        >;
+
+void add_device_conv3d_bwd_data_xdl_ndhwc_kzyxc_ndhwk_f16_instances(
+    std::vector<DeviceConvBwdDataPtr<PassThrough, PassThrough, PassThrough>>& instances)
+{
+    add_device_operation_instances(instances,
+                                   device_conv3d_bwd_data_xdl_ndhwc_kzyxc_ndhwk_f16_instances{});
+    add_device_operation_instances(
+        instances, device_conv3d_bwd_data_xdl_ndhwc_kzyxc_ndhwk_1x1_s1_p0_f16_instances{});
+}
+
+} // namespace device_conv2d_bwd_data_instance
+} // namespace device
+} // namespace tensor_operation
+} // namespace ck

library/src/tensor_operation_instance/gpu/convnd_bwd_data/device_conv3d_bwd_data_xdl_ndhwc_kzyxc_ndhwk_f32_instance.cpp

+#include <stdlib.h>
+#include "config.hpp"
+#include "device_convnd_bwd_data_xdl_ndhwc_kzyxc_ndhwk.hpp"
+#include "element_wise_operation.hpp"
+#include "device_operation_instance.hpp"
+
+namespace ck {
+namespace tensor_operation {
+namespace device {
+namespace device_conv2d_bwd_data_instance {
+
+using F32 = float;
+
+template <ck::index_t... Is>
+using S = ck::Sequence<Is...>;
+
+using PassThrough = ck::tensor_operation::element_wise::PassThrough;
+static constexpr auto ConvBwdDataDefault =
+    ck::tensor_operation::device::ConvolutionBackwardDataSpecialization_t::Default;
+
+static constexpr auto ConvBwdDataFilter1x1Stride1Pad0 =
+    ck::tensor_operation::device::ConvolutionBackwardDataSpecialization_t::Filter1x1Stride1Pad0;
+
+// Compilation parameters for in[n, hi, wi, c] * wei[k, y, x, c] = out[n, ho, wo, k]
+using device_conv3d_bwd_data_xdl_ndhwc_kzyxc_ndhwk_f32_instances =
+    std::tuple<
+        // clang-format off
+        //#############################################################################| InData| WeiData| OutData| AccData|          In|         Wei|         Out|        ConvBackward|    Num| Block|  MPer|  NPer| K0Per| K1| MPer| NPer| MXdl| NXdl|  ABlockTransfer| ABlockTransfer| ABlockTransfer| ABlockTransfer| ABlockTransfer| ABlockTransfer| ABlockLds|  BBlockTransfer| BBlockTransfer| BBlockTransfer| BlockTransfer| BBlockTransfer| BBlockTransfer| BBlockLds| CThreadTransfer| CThreadTransfer|
+        //#############################################################################|   Type|    Type|    Type|    Type| Elementwise| Elementwise| Elementwise|                Data|    Dim|  Size| Block| Block| Block|   |  XDL|  XDL|  Per|  Per|   ThreadCluster|  ThreadCluster| SrcAccessOrder|   SrcVectorDim|      SrcScalar|      DstScalar| AddExtraM|   ThreadCluster|  ThreadCluster| SrcAccessOrder|  SrcVectorDim|      SrcScalar|      DstScalar| AddExtraN| SrcDstVectorDim|       DstScalar|
+        //#############################################################################|       |        |        |        |   Operation|   Operation|   Operation|      Specialization|Spatial|      |      |      |      |   |     |     | Wave| Wave| Lengths_K0_M_K1|   ArrangeOrder|               |               |      PerVector|   PerVector_K1|          | Lengths_K0_N_K1|   ArrangeOrder|               |              |      PerVector|   PerVector_K1|          |                |       PerVector|
+        //#############################################################################|       |        |        |        |            |            |            |                    |       |      |      |      |      |   |     |     |     |     |                |               |               |               |               |               |          |                |               |               |              |               |               |          |                |                |
+        DeviceConvndBwdDataXdl_Input_N_Di_Hi_Wi_C_Weight_K_Z_Y_X_C_Output_N_Do_Ho_Wo_K<    F32,     F32,     F32,     F32, PassThrough, PassThrough, PassThrough,  ConvBwdDataDefault,   3,     256,   256,   128,     4,  4,   32,   32,    4,    2,     S<4, 64, 1>,     S<1, 0, 2>,     S<1, 0, 2>,              2,              4,              4,      true,     S<4, 64, 1>,     S<2, 0, 1>,     S<0, 2, 1>,             1,              2,              4,      true,               7,               1>,
+        DeviceConvndBwdDataXdl_Input_N_Di_Hi_Wi_C_Weight_K_Z_Y_X_C_Output_N_Do_Ho_Wo_K<    F32,     F32,     F32,     F32, PassThrough, PassThrough, PassThrough,  ConvBwdDataDefault,   3,     256,   128,   256,     4,  4,   32,   32,    2,    4,     S<4, 64, 1>,     S<1, 0, 2>,     S<1, 0, 2>,              2,              4,              4,      true,     S<4, 64, 1>,     S<2, 0, 1>,     S<0, 2, 1>,             1,              4,              4,      true,               7,               1>,
+        DeviceConvndBwdDataXdl_Input_N_Di_Hi_Wi_C_Weight_K_Z_Y_X_C_Output_N_Do_Ho_Wo_K<    F32,     F32,     F32,     F32, PassThrough, PassThrough, PassThrough,  ConvBwdDataDefault,   3,     128,   128,   128,     4,  4,   32,   32,    4,    2,     S<4, 32, 1>,     S<1, 0, 2>,     S<1, 0, 2>,              2,              4,              4,      true,     S<4, 32, 1>,     S<2, 0, 1>,     S<0, 2, 1>,             1,              4,              4,      true,               7,               1>,
+        DeviceConvndBwdDataXdl_Input_N_Di_Hi_Wi_C_Weight_K_Z_Y_X_C_Output_N_Do_Ho_Wo_K<    F32,     F32,     F32,     F32, PassThrough, PassThrough, PassThrough,  ConvBwdDataDefault,   3,     256,   128,   128,     4,  4,   32,   32,    2,    2,     S<4, 64, 1>,     S<1, 0, 2>,     S<1, 0, 2>,              2,              4,              4,      true,     S<4, 64, 1>,     S<2, 0, 1>,     S<0, 2, 1>,             1,              2,              4,      true,               7,               1>,
+        DeviceConvndBwdDataXdl_Input_N_Di_Hi_Wi_C_Weight_K_Z_Y_X_C_Output_N_Do_Ho_Wo_K<    F32,     F32,     F32,     F32, PassThrough, PassThrough, PassThrough,  ConvBwdDataDefault,   3,     128,   128,    64,     4,  4,   32,   32,    2,    2,     S<4, 32, 1>,     S<1, 0, 2>,     S<1, 0, 2>,              2,              4,              4,      true,     S<4, 32, 1>,     S<2, 0, 1>,     S<0, 2, 1>,             1,              2,              4,      true,               7,               1>,
+        DeviceConvndBwdDataXdl_Input_N_Di_Hi_Wi_C_Weight_K_Z_Y_X_C_Output_N_Do_Ho_Wo_K<    F32,     F32,     F32,     F32, PassThrough, PassThrough, PassThrough,  ConvBwdDataDefault,   3,     128,    64,   128,     4,  4,   32,   32,    2,    2,     S<4, 32, 1>,     S<1, 0, 2>,     S<1, 0, 2>,              2,              4,              4,      true,     S<4, 32, 1>,     S<2, 0, 1>,     S<0, 2, 1>,             1,              4,              4,      true,               7,               1>,
+        DeviceConvndBwdDataXdl_Input_N_Di_Hi_Wi_C_Weight_K_Z_Y_X_C_Output_N_Do_Ho_Wo_K<    F32,     F32,     F32,     F32, PassThrough, PassThrough, PassThrough,  ConvBwdDataDefault,   3,      64,    64,    64,     4,  4,   32,   32,    2,    2,     S<4, 16, 1>,     S<1, 0, 2>,     S<1, 0, 2>,              2,              4,              4,      true,     S<4, 16, 1>,     S<2, 0, 1>,     S<0, 2, 1>,             1,              4,              4,      true,               7,               1>,
+        DeviceConvndBwdDataXdl_Input_N_Di_Hi_Wi_C_Weight_K_Z_Y_X_C_Output_N_Do_Ho_Wo_K<    F32,     F32,     F32,     F32, PassThrough, PassThrough, PassThrough,  ConvBwdDataDefault,   3,     256,   128,    64,     4,  4,   32,   32,    2,    1,     S<4, 64, 1>,     S<1, 0, 2>,     S<1, 0, 2>,              2,              4,              4,      true,     S<4, 64, 1>,     S<2, 0, 1>,     S<0, 2, 1>,             1,              1,              4,      true,               7,               1>,
+        DeviceConvndBwdDataXdl_Input_N_Di_Hi_Wi_C_Weight_K_Z_Y_X_C_Output_N_Do_Ho_Wo_K<    F32,     F32,     F32,     F32, PassThrough, PassThrough, PassThrough,  ConvBwdDataDefault,   3,     256,    64,   128,     4,  4,   32,   32,    1,    2,     S<4, 64, 1>,     S<1, 0, 2>,     S<1, 0, 2>,              2,              4,              4,      true,     S<4, 64, 1>,     S<2, 0, 1>,     S<0, 2, 1>,             1,              2,              4,      true,               7,               1>,
+        DeviceConvndBwdDataXdl_Input_N_Di_Hi_Wi_C_Weight_K_Z_Y_X_C_Output_N_Do_Ho_Wo_K<    F32,     F32,     F32,     F32, PassThrough, PassThrough, PassThrough,  ConvBwdDataDefault,   3,     128,   128,    32,     4,  4,   32,   32,    2,    1,     S<4, 32, 1>,     S<1, 0, 2>,     S<1, 0, 2>,              2,              4,              4,      true,     S<4, 32, 1>,     S<2, 0, 1>,     S<0, 2, 1>,             1,              1,              4,      true,               7,               1>,
+        DeviceConvndBwdDataXdl_Input_N_Di_Hi_Wi_C_Weight_K_Z_Y_X_C_Output_N_Do_Ho_Wo_K<    F32,     F32,     F32,     F32, PassThrough, PassThrough, PassThrough,  ConvBwdDataDefault,   3,     128,    32,   128,     4,  4,   32,   32,    1,    2,     S<4, 32, 1>,     S<1, 0, 2>,     S<1, 0, 2>,              2,              4,              4,      true,     S<4, 32, 1>,     S<2, 0, 1>,     S<0, 2, 1>,             1,              4,              4,      true,               7,               1>,
+        DeviceConvndBwdDataXdl_Input_N_Di_Hi_Wi_C_Weight_K_Z_Y_X_C_Output_N_Do_Ho_Wo_K<    F32,     F32,     F32,     F32, PassThrough, PassThrough, PassThrough,  ConvBwdDataDefault,   3,      64,    64,    32,     4,  4,   32,   32,    2,    1,     S<4, 16, 1>,     S<1, 0, 2>,     S<1, 0, 2>,              2,              4,              4,      true,     S<4, 16, 1>,     S<2, 0, 1>,     S<0, 2, 1>,             1,              2,              4,      true,               7,               1>,
+        DeviceConvndBwdDataXdl_Input_N_Di_Hi_Wi_C_Weight_K_Z_Y_X_C_Output_N_Do_Ho_Wo_K<    F32,     F32,     F32,     F32, PassThrough, PassThrough, PassThrough,  ConvBwdDataDefault,   3,      64,    32,    64,     4,  4,   32,   32,    1,    2,     S<4, 16, 1>,     S<1, 0, 2>,     S<1, 0, 2>,              2,              4,              4,      true,     S<4, 16, 1>,     S<2, 0, 1>,     S<0, 2, 1>,             1,              4,              4,      true,               7,               1>
+        // clang-format on
+        >;
+
+using device_conv3d_bwd_data_xdl_ndhwc_kzyxc_ndhwk_1x1_s1_p0_f32_instances =
+    std::tuple<
+        // clang-format off
+        //#############################################################################| InData| WeiData| OutData| AccData|          In|         Wei|         Out|                     ConvBackward|    Num| Block|  MPer|  NPer| K0Per| K1| MPer| NPer| MXdl| NXdl|  ABlockTransfer| ABlockTransfer| ABlockTransfer| ABlockTransfer| ABlockTransfer| ABlockTransfer| ABlockLds|  BBlockTransfer| BBlockTransfer| BBlockTransfer| BlockTransfer| BBlockTransfer| BBlockTransfer| BBlockLds| CThreadTransfer| CThreadTransfer|
+        //#############################################################################|   Type|    Type|    Type|    Type| Elementwise| Elementwise| Elementwise|                             Data|    Dim|  Size| Block| Block| Block|   |  XDL|  XDL|  Per|  Per|   ThreadCluster|  ThreadCluster| SrcAccessOrder|   SrcVectorDim|      SrcScalar|      DstScalar| AddExtraM|   ThreadCluster|  ThreadCluster| SrcAccessOrder|  SrcVectorDim|      SrcScalar|      DstScalar| AddExtraN| SrcDstVectorDim|       DstScalar|
+        //#############################################################################|       |        |        |        |   Operation|   Operation|   Operation|                   Specialization|Spatial|      |      |      |      |   |     |     | Wave| Wave| Lengths_K0_M_K1|   ArrangeOrder|               |               |      PerVector|   PerVector_K1|          | Lengths_K0_N_K1|   ArrangeOrder|               |              |      PerVector|   PerVector_K1|          |                |       PerVector|
+        //#############################################################################|       |        |        |        |            |            |            |                                 |       |      |      |      |      |   |     |     |     |     |                |               |               |               |               |               |          |                |               |               |              |               |               |          |                |                |
+        DeviceConvndBwdDataXdl_Input_N_Di_Hi_Wi_C_Weight_K_Z_Y_X_C_Output_N_Do_Ho_Wo_K<    F32,     F32,     F32,     F32, PassThrough, PassThrough, PassThrough,  ConvBwdDataFilter1x1Stride1Pad0,   3,     256,   256,   128,     4,  4,   32,   32,    4,    2,     S<4, 64, 1>,     S<1, 0, 2>,     S<1, 0, 2>,              2,              4,              4,      true,     S<4, 64, 1>,     S<2, 0, 1>,     S<0, 2, 1>,             1,              2,              4,      true,               7,               1>,
+        DeviceConvndBwdDataXdl_Input_N_Di_Hi_Wi_C_Weight_K_Z_Y_X_C_Output_N_Do_Ho_Wo_K<    F32,     F32,     F32,     F32, PassThrough, PassThrough, PassThrough,  ConvBwdDataFilter1x1Stride1Pad0,   3,     256,   128,   256,     4,  4,   32,   32,    2,    4,     S<4, 64, 1>,     S<1, 0, 2>,     S<1, 0, 2>,              2,              4,              4,      true,     S<4, 64, 1>,     S<2, 0, 1>,     S<0, 2, 1>,             1,              4,              4,      true,               7,               1>,
+        DeviceConvndBwdDataXdl_Input_N_Di_Hi_Wi_C_Weight_K_Z_Y_X_C_Output_N_Do_Ho_Wo_K<    F32,     F32,     F32,     F32, PassThrough, PassThrough, PassThrough,  ConvBwdDataFilter1x1Stride1Pad0,   3,     128,   128,   128,     4,  4,   32,   32,    4,    2,     S<4, 32, 1>,     S<1, 0, 2>,     S<1, 0, 2>,              2,              4,              4,      true,     S<4, 32, 1>,     S<2, 0, 1>,     S<0, 2, 1>,             1,              4,              4,      true,               7,               1>,
+        DeviceConvndBwdDataXdl_Input_N_Di_Hi_Wi_C_Weight_K_Z_Y_X_C_Output_N_Do_Ho_Wo_K<    F32,     F32,     F32,     F32, PassThrough, PassThrough, PassThrough,  ConvBwdDataFilter1x1Stride1Pad0,   3,     256,   128,   128,     4,  4,   32,   32,    2,    2,     S<4, 64, 1>,     S<1, 0, 2>,     S<1, 0, 2>,              2,              4,              4,      true,     S<4, 64, 1>,     S<2, 0, 1>,     S<0, 2, 1>,             1,              2,              4,      true,               7,               1>,
+        DeviceConvndBwdDataXdl_Input_N_Di_Hi_Wi_C_Weight_K_Z_Y_X_C_Output_N_Do_Ho_Wo_K<    F32,     F32,     F32,     F32, PassThrough, PassThrough, PassThrough,  ConvBwdDataFilter1x1Stride1Pad0,   3,     128,   128,    64,     4,  4,   32,   32,    2,    2,     S<4, 32, 1>,     S<1, 0, 2>,     S<1, 0, 2>,              2,              4,              4,      true,     S<4, 32, 1>,     S<2, 0, 1>,     S<0, 2, 1>,             1,              2,              4,      true,               7,               1>,
+        DeviceConvndBwdDataXdl_Input_N_Di_Hi_Wi_C_Weight_K_Z_Y_X_C_Output_N_Do_Ho_Wo_K<    F32,     F32,     F32,     F32, PassThrough, PassThrough, PassThrough,  ConvBwdDataFilter1x1Stride1Pad0,   3,     128,    64,   128,     4,  4,   32,   32,    2,    2,     S<4, 32, 1>,     S<1, 0, 2>,     S<1, 0, 2>,              2,              4,              4,      true,     S<4, 32, 1>,     S<2, 0, 1>,     S<0, 2, 1>,             1,              4,              4,      true,               7,               1>,
+        DeviceConvndBwdDataXdl_Input_N_Di_Hi_Wi_C_Weight_K_Z_Y_X_C_Output_N_Do_Ho_Wo_K<    F32,     F32,     F32,     F32, PassThrough, PassThrough, PassThrough,  ConvBwdDataFilter1x1Stride1Pad0,   3,      64,    64,    64,     4,  4,   32,   32,    2,    2,     S<4, 16, 1>,     S<1, 0, 2>,     S<1, 0, 2>,              2,              4,              4,      true,     S<4, 16, 1>,     S<2, 0, 1>,     S<0, 2, 1>,             1,              4,              4,      true,               7,               1>,
+        DeviceConvndBwdDataXdl_Input_N_Di_Hi_Wi_C_Weight_K_Z_Y_X_C_Output_N_Do_Ho_Wo_K<    F32,     F32,     F32,     F32, PassThrough, PassThrough, PassThrough,  ConvBwdDataFilter1x1Stride1Pad0,   3,     256,   128,    64,     4,  4,   32,   32,    2,    1,     S<4, 64, 1>,     S<1, 0, 2>,     S<1, 0, 2>,              2,              4,              4,      true,     S<4, 64, 1>,     S<2, 0, 1>,     S<0, 2, 1>,             1,              1,              4,      true,               7,               1>,
+        DeviceConvndBwdDataXdl_Input_N_Di_Hi_Wi_C_Weight_K_Z_Y_X_C_Output_N_Do_Ho_Wo_K<    F32,     F32,     F32,     F32, PassThrough, PassThrough, PassThrough,  ConvBwdDataFilter1x1Stride1Pad0,   3,     256,    64,   128,     4,  4,   32,   32,    1,    2,     S<4, 64, 1>,     S<1, 0, 2>,     S<1, 0, 2>,              2,              4,              4,      true,     S<4, 64, 1>,     S<2, 0, 1>,     S<0, 2, 1>,             1,              2,              4,      true,               7,               1>,
+        DeviceConvndBwdDataXdl_Input_N_Di_Hi_Wi_C_Weight_K_Z_Y_X_C_Output_N_Do_Ho_Wo_K<    F32,     F32,     F32,     F32, PassThrough, PassThrough, PassThrough,  ConvBwdDataFilter1x1Stride1Pad0,   3,     128,   128,    32,     4,  4,   32,   32,    2,    1,     S<4, 32, 1>,     S<1, 0, 2>,     S<1, 0, 2>,              2,              4,              4,      true,     S<4, 32, 1>,     S<2, 0, 1>,     S<0, 2, 1>,             1,              1,              4,      true,               7,               1>,
+        DeviceConvndBwdDataXdl_Input_N_Di_Hi_Wi_C_Weight_K_Z_Y_X_C_Output_N_Do_Ho_Wo_K<    F32,     F32,     F32,     F32, PassThrough, PassThrough, PassThrough,  ConvBwdDataFilter1x1Stride1Pad0,   3,     128,    32,   128,     4,  4,   32,   32,    1,    2,     S<4, 32, 1>,     S<1, 0, 2>,     S<1, 0, 2>,              2,              4,              4,      true,     S<4, 32, 1>,     S<2, 0, 1>,     S<0, 2, 1>,             1,              4,              4,      true,               7,               1>,
+        DeviceConvndBwdDataXdl_Input_N_Di_Hi_Wi_C_Weight_K_Z_Y_X_C_Output_N_Do_Ho_Wo_K<    F32,     F32,     F32,     F32, PassThrough, PassThrough, PassThrough,  ConvBwdDataFilter1x1Stride1Pad0,   3,      64,    64,    32,     4,  4,   32,   32,    2,    1,     S<4, 16, 1>,     S<1, 0, 2>,     S<1, 0, 2>,              2,              4,              4,      true,     S<4, 16, 1>,     S<2, 0, 1>,     S<0, 2, 1>,             1,              2,              4,      true,               7,               1>,
+        DeviceConvndBwdDataXdl_Input_N_Di_Hi_Wi_C_Weight_K_Z_Y_X_C_Output_N_Do_Ho_Wo_K<    F32,     F32,     F32,     F32, PassThrough, PassThrough, PassThrough,  ConvBwdDataFilter1x1Stride1Pad0,   3,      64,    32,    64,     4,  4,   32,   32,    1,    2,     S<4, 16, 1>,     S<1, 0, 2>,     S<1, 0, 2>,              2,              4,              4,      true,     S<4, 16, 1>,     S<2, 0, 1>,     S<0, 2, 1>,             1,              4,              4,      true,               7,               1>
+        // clang-format on
+        >;
+
+void add_device_conv3d_bwd_data_xdl_ndhwc_kzyxc_ndhwk_f32_instances(
+    std::vector<DeviceConvBwdDataPtr<PassThrough, PassThrough, PassThrough>>& instances)
+{
+    add_device_operation_instances(instances,
+                                   device_conv3d_bwd_data_xdl_ndhwc_kzyxc_ndhwk_f32_instances{});
+    add_device_operation_instances(
+        instances, device_conv3d_bwd_data_xdl_ndhwc_kzyxc_ndhwk_1x1_s1_p0_f32_instances{});
+}
+
+} // namespace device_conv2d_bwd_data_instance
+} // namespace device
+} // namespace tensor_operation
+} // namespace ck

library/src/tensor_operation_instance/gpu/convnd_bwd_data/device_conv3d_bwd_data_xdl_ndhwc_kzyxc_ndhwk_int8_instance.cpp

+#include <stdlib.h>
+#include "config.hpp"
+#include "device_convnd_bwd_data_xdl_ndhwc_kzyxc_ndhwk.hpp"
+#include "element_wise_operation.hpp"
+#include "device_operation_instance.hpp"
+
+namespace ck {
+namespace tensor_operation {
+namespace device {
+namespace device_conv2d_bwd_data_instance {
+
+using DataType = int8_t;
+using AccType  = int32_t;
+
+template <ck::index_t... Is>
+using S = ck::Sequence<Is...>;
+
+using PassThrough = ck::tensor_operation::element_wise::PassThrough;
+static constexpr auto ConvBwdDataDefault =
+    ck::tensor_operation::device::ConvolutionBackwardDataSpecialization_t::Default;
+
+static constexpr auto ConvBwdDataFilter1x1Stride1Pad0 =
+    ck::tensor_operation::device::ConvolutionBackwardDataSpecialization_t::Filter1x1Stride1Pad0;
+
+// Compilation parameters for in[n, hi, wi, c] * wei[k, y, x, c] = out[n, ho, wo, k]
+using device_conv3d_bwd_data_xdl_ndhwc_kzyxc_ndhwk_int8_instances =
+    std::tuple<
+        // clang-format off
+        //#############################################################################|    InData|   WeiData|   OutData|    AccData|          In|         Wei|         Out|        ConvBackward|    Num| Block|  MPer|  NPer| K0Per|  K1| MPer| NPer| MXdl| NXdl|  ABlockTransfer| ABlockTransfer| ABlockTransfer| ABlockTransfer| ABlockTransfer| ABlockTransfer| ABlockLds|  BBlockTransfer| BBlockTransfer| BBlockTransfer| BlockTransfer| BBlockTransfer| BBlockTransfer| BBlockLds| CThreadTransfer| CThreadTransfer|
+        //#############################################################################|      Type|      Type|      Type|       Type| Elementwise| Elementwise| Elementwise|                Data|    Dim|  Size| Block| Block| Block|    |  XDL|  XDL|  Per|  Per|   ThreadCluster|  ThreadCluster| SrcAccessOrder|   SrcVectorDim|      SrcScalar|      DstScalar| AddExtraM|   ThreadCluster|  ThreadCluster| SrcAccessOrder|  SrcVectorDim|      SrcScalar|      DstScalar| AddExtraN| SrcDstVectorDim|       DstScalar|
+        //#############################################################################|          |          |          |           |   Operation|   Operation|   Operation|      Specialization|Spatial|      |      |      |      |    |     |     | Wave| Wave| Lengths_K0_M_K1|   ArrangeOrder|               |               |      PerVector|   PerVector_K1|          | Lengths_K0_N_K1|   ArrangeOrder|               |              |      PerVector|   PerVector_K1|          |                |       PerVector|
+        //#############################################################################|          |          |          |           |            |            |            |                    |       |      |      |      |      |    |     |     |     |     |                |               |               |               |               |               |          |                |               |               |              |               |               |          |                |                |
+        DeviceConvndBwdDataXdl_Input_N_Di_Hi_Wi_C_Weight_K_Z_Y_X_C_Output_N_Do_Ho_Wo_K<  DataType,  DataType,  DataType,    AccType, PassThrough, PassThrough, PassThrough,  ConvBwdDataDefault,   3,     256,   256,   128,     4,  16,   32,   32,    4,    2,     S<4, 64, 1>,     S<1, 0, 2>,     S<1, 0, 2>,              2,             16,             16,      true,     S<4, 64, 1>,     S<2, 0, 1>,     S<0, 2, 1>,             1,              2,             16,      true,               7,               1>,
+        DeviceConvndBwdDataXdl_Input_N_Di_Hi_Wi_C_Weight_K_Z_Y_X_C_Output_N_Do_Ho_Wo_K<  DataType,  DataType,  DataType,    AccType, PassThrough, PassThrough, PassThrough,  ConvBwdDataDefault,   3,     256,   128,   256,     4,  16,   32,   32,    2,    4,     S<4, 64, 1>,     S<1, 0, 2>,     S<1, 0, 2>,              2,             16,             16,      true,     S<4, 64, 1>,     S<2, 0, 1>,     S<0, 2, 1>,             1,              4,             16,      true,               7,               1>,
+        DeviceConvndBwdDataXdl_Input_N_Di_Hi_Wi_C_Weight_K_Z_Y_X_C_Output_N_Do_Ho_Wo_K<  DataType,  DataType,  DataType,    AccType, PassThrough, PassThrough, PassThrough,  ConvBwdDataDefault,   3,     128,   128,   128,     4,  16,   32,   32,    4,    2,     S<4, 32, 1>,     S<1, 0, 2>,     S<1, 0, 2>,              2,             16,             16,      true,     S<4, 32, 1>,     S<2, 0, 1>,     S<0, 2, 1>,             1,              2,             16,      true,               7,               1>,
+#if 1
+        DeviceConvndBwdDataXdl_Input_N_Di_Hi_Wi_C_Weight_K_Z_Y_X_C_Output_N_Do_Ho_Wo_K<  DataType,  DataType,  DataType,    AccType, PassThrough, PassThrough, PassThrough,  ConvBwdDataDefault,   3,     256,   128,   128,     4,  16,   32,   32,    2,    2,     S<4, 64, 1>,     S<1, 0, 2>,     S<1, 0, 2>,              2,             16,             16,      true,     S<4, 64, 1>,     S<2, 0, 1>,     S<0, 2, 1>,             1,              2,             16,      true,               7,               1>,
+        DeviceConvndBwdDataXdl_Input_N_Di_Hi_Wi_C_Weight_K_Z_Y_X_C_Output_N_Do_Ho_Wo_K<  DataType,  DataType,  DataType,    AccType, PassThrough, PassThrough, PassThrough,  ConvBwdDataDefault,   3,     128,    64,   128,     4,  16,   32,   32,    2,    2,     S<4, 32, 1>,     S<1, 0, 2>,     S<1, 0, 2>,              2,             16,             16,      true,     S<4, 32, 1>,     S<2, 0, 1>,     S<0, 2, 1>,             1,              4,             16,      true,               7,               1>,
+        DeviceConvndBwdDataXdl_Input_N_Di_Hi_Wi_C_Weight_K_Z_Y_X_C_Output_N_Do_Ho_Wo_K<  DataType,  DataType,  DataType,    AccType, PassThrough, PassThrough, PassThrough,  ConvBwdDataDefault,   3,     256,    64,   128,     4,  16,   32,   32,    1,    2,     S<4, 64, 1>,     S<1, 0, 2>,     S<1, 0, 2>,              2,             16,             16,      true,     S<4, 64, 1>,     S<2, 0, 1>,     S<0, 2, 1>,             1,              2,             16,      true,               7,               1>,
+        DeviceConvndBwdDataXdl_Input_N_Di_Hi_Wi_C_Weight_K_Z_Y_X_C_Output_N_Do_Ho_Wo_K<  DataType,  DataType,  DataType,    AccType, PassThrough, PassThrough, PassThrough,  ConvBwdDataDefault,   3,     128,    32,   128,     4,  16,   32,   32,    1,    2,     S<4, 32, 1>,     S<1, 0, 2>,     S<1, 0, 2>,              2,             16,             16,      true,     S<4, 32, 1>,     S<2, 0, 1>,     S<0, 2, 1>,             1,              4,             16,      true,               7,               1>,
+        DeviceConvndBwdDataXdl_Input_N_Di_Hi_Wi_C_Weight_K_Z_Y_X_C_Output_N_Do_Ho_Wo_K<  DataType,  DataType,  DataType,    AccType, PassThrough, PassThrough, PassThrough,  ConvBwdDataDefault,   3,      64,    32,    64,     4,  16,   32,   32,    1,    2,     S<4, 16, 1>,     S<1, 0, 2>,     S<1, 0, 2>,              2,             16,             16,      true,     S<4, 16, 1>,     S<2, 0, 1>,     S<0, 2, 1>,             1,              4,             16,      true,               7,               1>,
+#endif
+        DeviceConvndBwdDataXdl_Input_N_Di_Hi_Wi_C_Weight_K_Z_Y_X_C_Output_N_Do_Ho_Wo_K<  DataType,  DataType,  DataType,    AccType, PassThrough, PassThrough, PassThrough,  ConvBwdDataDefault,   3,     128,   128,    64,     4,  16,   32,   32,    2,    2,     S<4, 32, 1>,     S<1, 0, 2>,     S<1, 0, 2>,              2,             16,             16,      true,     S<4, 32, 1>,     S<2, 0, 1>,     S<0, 2, 1>,             1,              2,             16,      true,               7,               1>,
+        DeviceConvndBwdDataXdl_Input_N_Di_Hi_Wi_C_Weight_K_Z_Y_X_C_Output_N_Do_Ho_Wo_K<  DataType,  DataType,  DataType,    AccType, PassThrough, PassThrough, PassThrough,  ConvBwdDataDefault,   3,      64,    64,    64,     4,  16,   32,   32,    2,    2,     S<4, 16, 1>,     S<1, 0, 2>,     S<1, 0, 2>,              2,             16,             16,      true,     S<4, 16, 1>,     S<2, 0, 1>,     S<0, 2, 1>,             1,              4,             16,      true,               7,               1>,
+        DeviceConvndBwdDataXdl_Input_N_Di_Hi_Wi_C_Weight_K_Z_Y_X_C_Output_N_Do_Ho_Wo_K<  DataType,  DataType,  DataType,    AccType, PassThrough, PassThrough, PassThrough,  ConvBwdDataDefault,   3,     256,   128,    64,     4,  16,   32,   32,    2,    1,     S<4, 64, 1>,     S<1, 0, 2>,     S<1, 0, 2>,              2,             16,             16,      true,     S<4, 64, 1>,     S<2, 0, 1>,     S<0, 2, 1>,             1,              1,             16,      true,               7,               1>,
+        DeviceConvndBwdDataXdl_Input_N_Di_Hi_Wi_C_Weight_K_Z_Y_X_C_Output_N_Do_Ho_Wo_K<  DataType,  DataType,  DataType,    AccType, PassThrough, PassThrough, PassThrough,  ConvBwdDataDefault,   3,     128,   128,    32,     4,  16,   32,   32,    2,    1,     S<4, 32, 1>,     S<1, 0, 2>,     S<1, 0, 2>,              2,             16,             16,      true,     S<4, 32, 1>,     S<2, 0, 1>,     S<0, 2, 1>,             1,              1,             16,      true,               7,               1>,
+        DeviceConvndBwdDataXdl_Input_N_Di_Hi_Wi_C_Weight_K_Z_Y_X_C_Output_N_Do_Ho_Wo_K<  DataType,  DataType,  DataType,    AccType, PassThrough, PassThrough, PassThrough,  ConvBwdDataDefault,   3,      64,    64,    32,     4,  16,   32,   32,    2,    1,     S<4, 16, 1>,     S<1, 0, 2>,     S<1, 0, 2>,              2,             16,             16,      true,     S<4, 16, 1>,     S<2, 0, 1>,     S<0, 2, 1>,             1,              2,             16,      true,               7,               1>
+        // clang-format on
+        >;
+
+using device_conv3d_bwd_data_xdl_ndhwc_kzyxc_ndhwk_1x1_s1_p0_int8_instances =
+    std::tuple<
+        // clang-format off
+        //##############################################################################|    InData|   WeiData|   OutData|    AccData|          In|         Wei|         Out|                     ConvBackward|    Num| Block|  MPer|  NPer| K0Per|  K1| MPer| NPer| MXdl| NXdl|   ABlockTransfer| ABlockTransfer| ABlockTransfer| ABlockTransfer| ABlockTransfer| ABlockTransfer| ABlockLds|  BBlockTransfer| BBlockTransfer| BBlockTransfer| BlockTransfer| BBlockTransfer| BBlockTransfer| BBlockLds| CThreadTransfer| CThreadTransfer|
+        //##############################################################################|      Type|      Type|      Type|       Type| Elementwise| Elementwise| Elementwise|                             Data|    Dim|  Size| Block| Block| Block|    |  XDL|  XDL|  Per|  Per|    ThreadCluster|  ThreadCluster| SrcAccessOrder|   SrcVectorDim|      SrcScalar|      DstScalar| AddExtraM|   ThreadCluster|  ThreadCluster| SrcAccessOrder|  SrcVectorDim|      SrcScalar|      DstScalar| AddExtraN| SrcDstVectorDim|       DstScalar|
+        //##############################################################################|          |          |          |           |   Operation|   Operation|   Operation|                   Specialization|Spatial|      |      |      |      |    |     |     | Wave| Wave|  Lengths_K0_M_K1|   ArrangeOrder|               |               |      PerVector|   PerVector_K1|          | Lengths_K0_N_K1|   ArrangeOrder|               |              |      PerVector|   PerVector_K1|          |                |       PerVector|
+        //##############################################################################|          |          |          |           |            |            |            |                                 |       |      |      |      |      |    |     |     |     |     |                 |               |               |               |               |               |          |                |               |               |              |               |               |          |                |                |
+        DeviceConvndBwdDataXdl_Input_N_Di_Hi_Wi_C_Weight_K_Z_Y_X_C_Output_N_Do_Ho_Wo_K<   DataType,  DataType,  DataType,    AccType, PassThrough, PassThrough, PassThrough,  ConvBwdDataFilter1x1Stride1Pad0,   3,     256,   256,   128,     4,  16,   32,   32,    4,    2,      S<4, 64, 1>,     S<1, 0, 2>,     S<1, 0, 2>,              2,             16,             16,      true,     S<4, 64, 1>,     S<2, 0, 1>,     S<0, 2, 1>,             1,              2,             16,      true,               7,               1>,
+        DeviceConvndBwdDataXdl_Input_N_Di_Hi_Wi_C_Weight_K_Z_Y_X_C_Output_N_Do_Ho_Wo_K<   DataType,  DataType,  DataType,    AccType, PassThrough, PassThrough, PassThrough,  ConvBwdDataFilter1x1Stride1Pad0,   3,     256,   128,   256,     4,  16,   32,   32,    2,    4,      S<4, 64, 1>,     S<1, 0, 2>,     S<1, 0, 2>,              2,             16,             16,      true,     S<4, 64, 1>,     S<2, 0, 1>,     S<0, 2, 1>,             1,              4,             16,      true,               7,               1>,
+        DeviceConvndBwdDataXdl_Input_N_Di_Hi_Wi_C_Weight_K_Z_Y_X_C_Output_N_Do_Ho_Wo_K<   DataType,  DataType,  DataType,    AccType, PassThrough, PassThrough, PassThrough,  ConvBwdDataFilter1x1Stride1Pad0,   3,     128,   128,   128,     4,  16,   32,   32,    4,    2,      S<4, 32, 1>,     S<1, 0, 2>,     S<1, 0, 2>,              2,             16,             16,      true,     S<4, 32, 1>,     S<2, 0, 1>,     S<0, 2, 1>,             1,              2,             16,      true,               7,               1>,
+        DeviceConvndBwdDataXdl_Input_N_Di_Hi_Wi_C_Weight_K_Z_Y_X_C_Output_N_Do_Ho_Wo_K<   DataType,  DataType,  DataType,    AccType, PassThrough, PassThrough, PassThrough,  ConvBwdDataFilter1x1Stride1Pad0,   3,     256,   128,   128,     4,  16,   32,   32,    2,    2,      S<4, 64, 1>,     S<1, 0, 2>,     S<1, 0, 2>,              2,             16,             16,      true,     S<4, 64, 1>,     S<2, 0, 1>,     S<0, 2, 1>,             1,              2,             16,      true,               7,               1>,
+        DeviceConvndBwdDataXdl_Input_N_Di_Hi_Wi_C_Weight_K_Z_Y_X_C_Output_N_Do_Ho_Wo_K<   DataType,  DataType,  DataType,    AccType, PassThrough, PassThrough, PassThrough,  ConvBwdDataFilter1x1Stride1Pad0,   3,     128,   128,    64,     4,  16,   32,   32,    2,    2,      S<4, 32, 1>,     S<1, 0, 2>,     S<1, 0, 2>,              2,             16,             16,      true,     S<4, 32, 1>,     S<2, 0, 1>,     S<0, 2, 1>,             1,              2,             16,      true,               7,               1>,
+        DeviceConvndBwdDataXdl_Input_N_Di_Hi_Wi_C_Weight_K_Z_Y_X_C_Output_N_Do_Ho_Wo_K<   DataType,  DataType,  DataType,    AccType, PassThrough, PassThrough, PassThrough,  ConvBwdDataFilter1x1Stride1Pad0,   3,     128,    64,   128,     4,  16,   32,   32,    2,    2,      S<4, 32, 1>,     S<1, 0, 2>,     S<1, 0, 2>,              2,             16,             16,      true,     S<4, 32, 1>,     S<2, 0, 1>,     S<0, 2, 1>,             1,              4,             16,      true,               7,               1>,
+        DeviceConvndBwdDataXdl_Input_N_Di_Hi_Wi_C_Weight_K_Z_Y_X_C_Output_N_Do_Ho_Wo_K<   DataType,  DataType,  DataType,    AccType, PassThrough, PassThrough, PassThrough,  ConvBwdDataFilter1x1Stride1Pad0,   3,      64,    64,    64,     4,  16,   32,   32,    2,    2,      S<4, 16, 1>,     S<1, 0, 2>,     S<1, 0, 2>,              2,             16,             16,      true,     S<4, 16, 1>,     S<2, 0, 1>,     S<0, 2, 1>,             1,              4,             16,      true,               7,               1>,
+        DeviceConvndBwdDataXdl_Input_N_Di_Hi_Wi_C_Weight_K_Z_Y_X_C_Output_N_Do_Ho_Wo_K<   DataType,  DataType,  DataType,    AccType, PassThrough, PassThrough, PassThrough,  ConvBwdDataFilter1x1Stride1Pad0,   3,     256,   128,    64,     4,  16,   32,   32,    2,    1,      S<4, 64, 1>,     S<1, 0, 2>,     S<1, 0, 2>,              2,             16,             16,      true,     S<4, 64, 1>,     S<2, 0, 1>,     S<0, 2, 1>,             1,              1,             16,      true,               7,               1>,
+        DeviceConvndBwdDataXdl_Input_N_Di_Hi_Wi_C_Weight_K_Z_Y_X_C_Output_N_Do_Ho_Wo_K<   DataType,  DataType,  DataType,    AccType, PassThrough, PassThrough, PassThrough,  ConvBwdDataFilter1x1Stride1Pad0,   3,     256,    64,   128,     4,  16,   32,   32,    1,    2,      S<4, 64, 1>,     S<1, 0, 2>,     S<1, 0, 2>,              2,             16,             16,      true,     S<4, 64, 1>,     S<2, 0, 1>,     S<0, 2, 1>,             1,              2,             16,      true,               7,               1>,
+        DeviceConvndBwdDataXdl_Input_N_Di_Hi_Wi_C_Weight_K_Z_Y_X_C_Output_N_Do_Ho_Wo_K<   DataType,  DataType,  DataType,    AccType, PassThrough, PassThrough, PassThrough,  ConvBwdDataFilter1x1Stride1Pad0,   3,     128,   128,    32,     4,  16,   32,   32,    2,    1,      S<4, 32, 1>,     S<1, 0, 2>,     S<1, 0, 2>,              2,             16,             16,      true,     S<4, 32, 1>,     S<2, 0, 1>,     S<0, 2, 1>,             1,              1,             16,      true,               7,               1>,
+        DeviceConvndBwdDataXdl_Input_N_Di_Hi_Wi_C_Weight_K_Z_Y_X_C_Output_N_Do_Ho_Wo_K<   DataType,  DataType,  DataType,    AccType, PassThrough, PassThrough, PassThrough,  ConvBwdDataFilter1x1Stride1Pad0,   3,     128,    32,   128,     4,  16,   32,   32,    1,    2,      S<4, 32, 1>,     S<1, 0, 2>,     S<1, 0, 2>,              2,             16,             16,      true,     S<4, 32, 1>,     S<2, 0, 1>,     S<0, 2, 1>,             1,              4,             16,      true,               7,               1>,
+        DeviceConvndBwdDataXdl_Input_N_Di_Hi_Wi_C_Weight_K_Z_Y_X_C_Output_N_Do_Ho_Wo_K<   DataType,  DataType,  DataType,    AccType, PassThrough, PassThrough, PassThrough,  ConvBwdDataFilter1x1Stride1Pad0,   3,      64,    64,    32,     4,  16,   32,   32,    2,    1,      S<4, 16, 1>,     S<1, 0, 2>,     S<1, 0, 2>,              2,             16,             16,      true,     S<4, 16, 1>,     S<2, 0, 1>,     S<0, 2, 1>,             1,              2,             16,      true,               7,               1>,
+        DeviceConvndBwdDataXdl_Input_N_Di_Hi_Wi_C_Weight_K_Z_Y_X_C_Output_N_Do_Ho_Wo_K<   DataType,  DataType,  DataType,    AccType, PassThrough, PassThrough, PassThrough,  ConvBwdDataFilter1x1Stride1Pad0,   3,      64,    32,    64,     4,  16,   32,   32,    1,    2,      S<4, 16, 1>,     S<1, 0, 2>,     S<1, 0, 2>,              2,             16,             16,      true,     S<4, 16, 1>,     S<2, 0, 1>,     S<0, 2, 1>,             1,              4,             16,      true,               7,               1>
+        // clang-format on
+        >;
+
+void add_device_conv3d_bwd_data_xdl_ndhwc_kzyxc_ndhwk_int8_instances(
+    std::vector<DeviceConvBwdDataPtr<PassThrough, PassThrough, PassThrough>>& instances)
+{
+    add_device_operation_instances(instances,
+                                   device_conv3d_bwd_data_xdl_ndhwc_kzyxc_ndhwk_int8_instances{});
+    add_device_operation_instances(
+        instances, device_conv3d_bwd_data_xdl_ndhwc_kzyxc_ndhwk_1x1_s1_p0_int8_instances{});
+}
+
+} // namespace device_conv2d_bwd_data_instance
+} // namespace device
+} // namespace tensor_operation
+} // namespace ck

profiler/CMakeLists.txt

@@ -32,7 +32,7 @@ set(PROFILER_SOURCE
     src/profile_conv_fwd_bias_relu.cpp
     src/profile_conv_fwd_bias_relu_add.cpp
     src/profile_conv_fwd_bias_relu_atomic_add.cpp
-    src/profile_conv_bwd_data.cpp
+    src/profile_convnd_bwd_data.cpp
     src/profile_reduce.cpp
     src/profile_grouped_gemm.cpp
 )
@@ -50,7 +50,7 @@ target_link_libraries(ckProfiler PRIVATE device_conv2d_fwd_instance)
 target_link_libraries(ckProfiler PRIVATE device_conv2d_fwd_bias_relu_instance)
 target_link_libraries(ckProfiler PRIVATE device_conv2d_fwd_bias_relu_add_instance)
 target_link_libraries(ckProfiler PRIVATE device_conv2d_fwd_bias_relu_atomic_add_instance)
-target_link_libraries(ckProfiler PRIVATE device_conv2d_bwd_data_instance)
+target_link_libraries(ckProfiler PRIVATE device_convnd_bwd_data_instance)
 target_link_libraries(ckProfiler PRIVATE device_reduce_instance)
 target_link_libraries(ckProfiler PRIVATE device_reduce_instance)
 target_link_libraries(ckProfiler PRIVATE device_grouped_gemm_instance)

profiler/include/profile_conv_bwd_data_impl.hpp

@@ -42,6 +42,7 @@ template <int NDimSpatial,
           typename InDataType,
           typename WeiDataType,
           typename OutDataType,
+          typename AccDataType,
           typename InLayout,
           typename WeiLayout,
           typename OutLayout>
@@ -123,6 +124,7 @@ void profile_conv_bwd_data_impl(int do_verification,
             ck::tensor_operation::host::ReferenceConvBwdData<InDataType,
                                                              WeiDataType,
                                                              OutDataType,
+                                                             AccDataType,
                                                              InElementOp,
                                                              WeiElementOp,
                                                              OutElementOp>;

profiler/include/profile_convnd_bwd_data_impl.hpp

+#pragma once
+#include "config.hpp"
+#include "device.hpp"
+#include "conv_utils.hpp"
+#include "host_tensor.hpp"
+#include "host_tensor_generator.hpp"
+#include "tensor_layout.hpp"
+#include "device_tensor.hpp"
+#include "device_conv_bwd_data.hpp"
+#include "element_wise_operation.hpp"
+#include "reference_conv_bwd_data.hpp"
+
+using F16  = ck::half_t;
+using F32  = float;
+using BF16 = ushort;
+using INT8 = int8_t;
+namespace ck {
+namespace tensor_operation {
+namespace device {
+namespace device_conv2d_bwd_data_instance {
+
+using DeviceConvBwdDataNoOpPtr =
+    DeviceConvBwdDataPtr<ck::tensor_operation::element_wise::PassThrough,
+                         ck::tensor_operation::element_wise::PassThrough,
+                         ck::tensor_operation::element_wise::PassThrough>;
+void add_device_conv1d_bwd_data_xdl_nwc_kxc_nwk_f32_instances(
+    std::vector<DeviceConvBwdDataNoOpPtr>&);
+void add_device_conv1d_bwd_data_xdl_nwc_kxc_nwk_f16_instances(
+    std::vector<DeviceConvBwdDataNoOpPtr>&);
+void add_device_conv1d_bwd_data_xdl_nwc_kxc_nwk_bf16_instances(
+    std::vector<DeviceConvBwdDataNoOpPtr>&);
+void add_device_conv1d_bwd_data_xdl_nwc_kxc_nwk_int8_instances(
+    std::vector<DeviceConvBwdDataNoOpPtr>&);
+
+void add_device_conv2d_bwd_data_xdl_nhwc_kyxc_nhwk_f32_instances(
+    std::vector<DeviceConvBwdDataNoOpPtr>&);
+void add_device_conv2d_bwd_data_xdl_nhwc_kyxc_nhwk_f16_instances(
+    std::vector<DeviceConvBwdDataNoOpPtr>&);
+void add_device_conv2d_bwd_data_xdl_nhwc_kyxc_nhwk_bf16_instances(
+    std::vector<DeviceConvBwdDataNoOpPtr>&);
+void add_device_conv2d_bwd_data_xdl_nhwc_kyxc_nhwk_int8_instances(
+    std::vector<DeviceConvBwdDataNoOpPtr>&);
+
+void add_device_conv3d_bwd_data_xdl_ndhwc_kzyxc_ndhwk_f32_instances(
+    std::vector<DeviceConvBwdDataNoOpPtr>&);
+void add_device_conv3d_bwd_data_xdl_ndhwc_kzyxc_ndhwk_f16_instances(
+    std::vector<DeviceConvBwdDataNoOpPtr>&);
+void add_device_conv3d_bwd_data_xdl_ndhwc_kzyxc_ndhwk_bf16_instances(
+    std::vector<DeviceConvBwdDataNoOpPtr>&);
+void add_device_conv3d_bwd_data_xdl_ndhwc_kzyxc_ndhwk_int8_instances(
+    std::vector<DeviceConvBwdDataNoOpPtr>&);
+} // namespace device_conv2d_bwd_data_instance
+} // namespace device
+} // namespace tensor_operation
+} // namespace ck
+
+namespace ck {
+namespace profiler {
+using DeviceConvBwdDataNoOpPtr =
+    ck::tensor_operation::device::device_conv2d_bwd_data_instance::DeviceConvBwdDataNoOpPtr;
+
+template <typename InLayout>
+HostTensorDescriptor get_input_host_tensor_descriptor(const std::vector<std::size_t>& dims,
+                                                      int num_dim_spatial = 2)
+{
+    namespace tl = ck::tensor_layout::convolution;
+
+    switch(num_dim_spatial)
+    {
+    case 3: {
+        return ck::conv_util::GetHostTensorDescriptor(dims, InLayout{});
+    }
+    case 2: {
+        return ck::conv_util::GetHostTensorDescriptor(dims, InLayout{});
+    }
+    case 1: {
+        return ck::conv_util::GetHostTensorDescriptor(dims, InLayout{});
+    }
+    default: {
+        throw std::runtime_error("Unsupported number of spatial dimensions provided!");
+    }
+    }
+}
+template <typename WeiLayout>
+HostTensorDescriptor get_filters_host_tensor_descriptor(const std::vector<std::size_t>& dims,
+                                                        int num_dim_spatial = 2)
+{
+    namespace tl = ck::tensor_layout::convolution;
+
+    switch(num_dim_spatial)
+    {
+    case 3: {
+        return ck::conv_util::GetHostTensorDescriptor(dims, WeiLayout{});
+    }
+    case 2: {
+        return ck::conv_util::GetHostTensorDescriptor(dims, WeiLayout{});
+    }
+    case 1: {
+        return ck::conv_util::GetHostTensorDescriptor(dims, WeiLayout{});
+    }
+    default: {
+        throw std::runtime_error("Unsupported number of spatial dimensions provided!");
+    }
+    }
+}
+template <typename OutLayout>
+HostTensorDescriptor get_output_host_ensor_descriptor(const std::vector<std::size_t>& dims,
+                                                      int num_dim_spatial = 2)
+{
+    namespace tl = ck::tensor_layout::convolution;
+
+    switch(num_dim_spatial)
+    {
+    case 3: {
+        return ck::conv_util::GetHostTensorDescriptor(dims, OutLayout{});
+    }
+    case 2: {
+        return ck::conv_util::GetHostTensorDescriptor(dims, OutLayout{});
+    }
+    case 1: {
+        return ck::conv_util::GetHostTensorDescriptor(dims, OutLayout{});
+    }
+
+    default: {
+        throw std::runtime_error("Unsupported number of spatial dimensions provided!");
+    }
+    }
+}
+template <typename InDataType, typename WeiDataType, typename OutDataType>
+void get_device_conv_bwd_data_op_ptr(
+    InDataType, WeiDataType, OutDataType, std::vector<DeviceConvBwdDataNoOpPtr>&, int)
+{
+    std::cout << "can not find device conv bwd data" << std::endl;
+    exit(1);
+}
+template <>
+void get_device_conv_bwd_data_op_ptr(
+    F32, F32, F32, std::vector<DeviceConvBwdDataNoOpPtr>& conv_ptrs, int num_dim_spatial)
+{
+    switch(num_dim_spatial)
+    {
+    case 1:
+        ck::tensor_operation::device::device_conv2d_bwd_data_instance::
+            add_device_conv1d_bwd_data_xdl_nwc_kxc_nwk_f32_instances(conv_ptrs);
+        break;
+    case 2:
+        ck::tensor_operation::device::device_conv2d_bwd_data_instance::
+            add_device_conv2d_bwd_data_xdl_nhwc_kyxc_nhwk_f32_instances(conv_ptrs);
+        break;
+    case 3:
+        ck::tensor_operation::device::device_conv2d_bwd_data_instance::
+            add_device_conv3d_bwd_data_xdl_ndhwc_kzyxc_ndhwk_f32_instances(conv_ptrs);
+        break;
+    default: break;
+    }
+}
+template <>
+void get_device_conv_bwd_data_op_ptr(
+    F16, F16, F16, std::vector<DeviceConvBwdDataNoOpPtr>& conv_ptrs, int num_dim_spatial)
+{
+    switch(num_dim_spatial)
+    {
+    case 1:
+        ck::tensor_operation::device::device_conv2d_bwd_data_instance::
+            add_device_conv1d_bwd_data_xdl_nwc_kxc_nwk_f16_instances(conv_ptrs);
+        break;
+    case 2:
+        ck::tensor_operation::device::device_conv2d_bwd_data_instance::
+            add_device_conv2d_bwd_data_xdl_nhwc_kyxc_nhwk_f16_instances(conv_ptrs);
+        break;
+    case 3:
+        ck::tensor_operation::device::device_conv2d_bwd_data_instance::
+            add_device_conv3d_bwd_data_xdl_ndhwc_kzyxc_ndhwk_f16_instances(conv_ptrs);
+        break;
+    default: break;
+    }
+}
+template <>
+void get_device_conv_bwd_data_op_ptr(
+    BF16, BF16, BF16, std::vector<DeviceConvBwdDataNoOpPtr>& conv_ptrs, int num_dim_spatial)
+{
+    switch(num_dim_spatial)
+    {
+    case 1:
+        ck::tensor_operation::device::device_conv2d_bwd_data_instance::
+            add_device_conv1d_bwd_data_xdl_nwc_kxc_nwk_bf16_instances(conv_ptrs);
+        break;
+    case 2:
+        ck::tensor_operation::device::device_conv2d_bwd_data_instance::
+            add_device_conv2d_bwd_data_xdl_nhwc_kyxc_nhwk_bf16_instances(conv_ptrs);
+        break;
+    case 3:
+        ck::tensor_operation::device::device_conv2d_bwd_data_instance::
+            add_device_conv3d_bwd_data_xdl_ndhwc_kzyxc_ndhwk_bf16_instances(conv_ptrs);
+        break;
+    default: break;
+    }
+}
+template <>
+void get_device_conv_bwd_data_op_ptr(
+    INT8, INT8, INT8, std::vector<DeviceConvBwdDataNoOpPtr>& conv_ptrs, int num_dim_spatial)
+{
+    switch(num_dim_spatial)
+    {
+    case 1:
+        ck::tensor_operation::device::device_conv2d_bwd_data_instance::
+            add_device_conv1d_bwd_data_xdl_nwc_kxc_nwk_int8_instances(conv_ptrs);
+        break;
+    case 2:
+        ck::tensor_operation::device::device_conv2d_bwd_data_instance::
+            add_device_conv2d_bwd_data_xdl_nhwc_kyxc_nhwk_int8_instances(conv_ptrs);
+        break;
+    case 3:
+        ck::tensor_operation::device::device_conv2d_bwd_data_instance::
+            add_device_conv3d_bwd_data_xdl_ndhwc_kzyxc_ndhwk_int8_instances(conv_ptrs);
+        break;
+    default: break;
+    }
+}
+
+template <typename T>
+static bool check_out(const Tensor<T>& ref, const Tensor<T>& result)
+{
+    float max_diff = 1e-6;
+
+    for(int i = 0; i < ref.mData.size(); ++i)
+    {
+        float diff = std::abs(double(ref.mData[i]) - double(result.mData[i]));
+        if(max_diff < diff)
+        {
+            return false;
+        }
+    }
+    return true;
+}
+template <typename DataType>
+void show_data_nhwc_layout(Tensor<DataType>& nhwc)
+{
+    std::cout << "[";
+    for(int n = 0; n < nhwc.mDesc.GetLengths()[0]; n++)
+    {
+        std::cout << "[";
+        for(int hi = 0; hi < nhwc.mDesc.GetLengths()[2]; hi++)
+        {
+            std::cout << "[";
+            for(int wi = 0; wi < nhwc.mDesc.GetLengths()[3]; wi++)
+            {
+                std::cout << "[";
+                for(int c = 0; c < nhwc.mDesc.GetLengths()[1]; c++)
+                {
+                    std::cout << static_cast<float>(nhwc(n, c, hi, wi)) << "  ";
+                }
+                std::cout << "]";
+            }
+            std::cout << "]";
+        }
+        std::cout << "]";
+    }
+    std::cout << "]";
+}
+
+template <int NDimSpatial,
+          typename InDataType,
+          typename WeiDataType,
+          typename OutDataType,
+          typename AccDataType,
+          typename InLayout,
+          typename WeiLayout,
+          typename OutLayout>
+bool profile_convnd_bwd_data_impl(int do_verification,
+                                  int init_method,
+                                  bool do_log,
+                                  int nrepeat,
+                                  ck::index_t N,
+                                  ck::index_t K,
+                                  ck::index_t C,
+                                  std::vector<ck::index_t> input_spatial_lengths,
+                                  std::vector<ck::index_t> filter_spatial_lengths,
+                                  std::vector<ck::index_t> output_spatial_lengths,
+                                  std::vector<ck::index_t> conv_filter_strides,
+                                  std::vector<ck::index_t> conv_filter_dilations,
+                                  std::vector<ck::index_t> input_left_pads,
+                                  std::vector<ck::index_t> input_right_pads)
+{
+    using InElementOp  = ck::tensor_operation::element_wise::PassThrough;
+    using WeiElementOp = ck::tensor_operation::element_wise::PassThrough;
+    using OutElementOp = ck::tensor_operation::element_wise::PassThrough;
+
+    const auto in_element_op  = InElementOp{};
+    const auto wei_element_op = WeiElementOp{};
+    const auto out_element_op = OutElementOp{};
+
+    std::vector<std::size_t> input_dims{static_cast<std::size_t>(N), static_cast<std::size_t>(C)};
+    input_dims.insert(
+        std::end(input_dims), std::begin(input_spatial_lengths), std::end(input_spatial_lengths));
+
+    std::vector<std::size_t> filter_dims{static_cast<std::size_t>(K), static_cast<std::size_t>(C)};
+    filter_dims.insert(std::end(filter_dims),
+                       std::begin(filter_spatial_lengths),
+                       std::end(filter_spatial_lengths));
+
+    std::vector<std::size_t> output_dims{static_cast<std::size_t>(N), static_cast<std::size_t>(K)};
+    output_dims.insert(std::end(output_dims),
+                       std::begin(output_spatial_lengths),
+                       std::end(output_spatial_lengths));
+
+    Tensor<InDataType> in_n_c_hi_wi_host_result(
+        get_input_host_tensor_descriptor<InLayout>(input_dims, NDimSpatial));
+    Tensor<InDataType> in_n_c_hi_wi_device_result(
+        get_input_host_tensor_descriptor<InLayout>(input_dims, NDimSpatial));
+    Tensor<WeiDataType> wei_k_c_y_x(
+        get_filters_host_tensor_descriptor<WeiLayout>(filter_dims, NDimSpatial));
+    Tensor<OutDataType> out_n_k_ho_wo(
+        get_output_host_ensor_descriptor<OutLayout>(output_dims, NDimSpatial));
+
+    std::cout << "in_n_c_hi_wi: " << in_n_c_hi_wi_host_result.mDesc << std::endl;
+    std::cout << "wei_k_c_y_x: " << wei_k_c_y_x.mDesc << std::endl;
+    std::cout << "out_n_k_ho_wo: " << out_n_k_ho_wo.mDesc << std::endl;
+
+    switch(init_method)
+    {
+    case 0: break;
+    case 1:
+        out_n_k_ho_wo.GenerateTensorValue(GeneratorTensor_2<OutDataType>{-5, 5});
+        wei_k_c_y_x.GenerateTensorValue(GeneratorTensor_2<WeiDataType>{-5, 5});
+        break;
+    default:
+        out_n_k_ho_wo.GenerateTensorValue(GeneratorTensor_1<OutDataType>{1});
+        wei_k_c_y_x.GenerateTensorValue(GeneratorTensor_1<WeiDataType>{1});
+    }
+
+    DeviceMem in_device_buf(sizeof(InDataType) *
+                            in_n_c_hi_wi_device_result.mDesc.GetElementSpace());
+    DeviceMem wei_device_buf(sizeof(WeiDataType) * wei_k_c_y_x.mDesc.GetElementSpace());
+    DeviceMem out_device_buf(sizeof(OutDataType) * out_n_k_ho_wo.mDesc.GetElementSpace());
+
+    out_device_buf.ToDevice(out_n_k_ho_wo.mData.data());
+    wei_device_buf.ToDevice(wei_k_c_y_x.mData.data());
+
+    // reset input to zero
+    in_n_c_hi_wi_device_result.GenerateTensorValue(GeneratorTensor_1<InDataType>{0});
+    in_device_buf.ToDevice(in_n_c_hi_wi_device_result.mData.data());
+
+    if(do_verification)
+    {
+        auto RunReference = [&](auto& ref_conv) {
+            auto ref_invoker = ref_conv.MakeInvoker();
+
+            auto ref_argument = ref_conv.MakeArgument(in_n_c_hi_wi_host_result,
+                                                      wei_k_c_y_x,
+                                                      out_n_k_ho_wo,
+                                                      conv_filter_strides,
+                                                      conv_filter_dilations,
+                                                      input_left_pads,
+                                                      input_right_pads,
+                                                      InElementOp{},
+                                                      WeiElementOp{},
+                                                      OutElementOp{});
+            ref_invoker.Run(ref_argument);
+        };
+        switch(NDimSpatial)
+        {
+        case 3: {
+            auto ref_conv = ck::tensor_operation::host::ReferenceConvBwdData<InDataType,
+                                                                             WeiDataType,
+                                                                             OutDataType,
+                                                                             AccDataType,
+                                                                             InElementOp,
+                                                                             WeiElementOp,
+                                                                             OutElementOp,
+                                                                             3>();
+            RunReference(ref_conv);
+            break;
+        }
+        case 2: {
+            auto ref_conv = ck::tensor_operation::host::ReferenceConvBwdData<InDataType,
+                                                                             WeiDataType,
+                                                                             OutDataType,
+                                                                             AccDataType,
+                                                                             InElementOp,
+                                                                             WeiElementOp,
+                                                                             OutElementOp,
+                                                                             2>();
+            RunReference(ref_conv);
+            break;
+        }
+        case 1: {
+            auto ref_conv = ck::tensor_operation::host::ReferenceConvBwdData<InDataType,
+                                                                             WeiDataType,
+                                                                             OutDataType,
+                                                                             AccDataType,
+                                                                             InElementOp,
+                                                                             WeiElementOp,
+                                                                             OutElementOp,
+                                                                             1>();
+            RunReference(ref_conv);
+            break;
+        }
+        default: {
+            throw std::runtime_error("Unsupported number of spatial dimensions provided!");
+        }
+        }
+    }
+
+    // add device Conv instances
+    std::vector<DeviceConvBwdDataNoOpPtr> conv_ptrs;
+    get_device_conv_bwd_data_op_ptr(
+        InDataType{}, WeiDataType{}, OutDataType{}, conv_ptrs, NDimSpatial);
+
+    if(conv_ptrs.size() <= 0)
+    {
+        throw std::runtime_error("wrong! no device Conv instance found");
+    }
+
+    std::string best_conv_name;
+    float best_ave_time   = 0;
+    float best_tflops     = 0;
+    float best_gb_per_sec = 0;
+
+    // profile device Conv instances
+    bool success = true;
+    for(auto& conv_ptr : conv_ptrs)
+    {
+        auto argument_ptr = conv_ptr->MakeArgumentPointer(
+            static_cast<InDataType*>(in_device_buf.GetDeviceBuffer()),
+            static_cast<WeiDataType*>(wei_device_buf.GetDeviceBuffer()),
+            static_cast<OutDataType*>(out_device_buf.GetDeviceBuffer()),
+            N,
+            K,
+            C,
+            input_spatial_lengths,
+            filter_spatial_lengths,
+            output_spatial_lengths,
+            conv_filter_strides,
+            conv_filter_dilations,
+            input_left_pads,
+            input_right_pads,
+            in_element_op,
+            wei_element_op,
+            out_element_op);
+
+        auto invoker_ptr = conv_ptr->MakeInvokerPointer();
+
+        if(conv_ptr->IsSupportedArgument(argument_ptr.get()))
+        {
+            std::string conv_name = conv_ptr->GetTypeString();
+
+            float ave_time = invoker_ptr->Run(argument_ptr.get(), nrepeat);
+
+            std::size_t flop =
+                ck::conv_util::GetFlops(N, C, K, filter_spatial_lengths, output_spatial_lengths);
+            std::size_t num_btype = ck::conv_util::GetBtype<InDataType, WeiDataType, OutDataType>(
+                N, C, K, input_spatial_lengths, filter_spatial_lengths, output_spatial_lengths);
+
+            float tflops     = static_cast<float>(flop) / 1.E9 / ave_time;
+            float gb_per_sec = num_btype / 1.E6 / ave_time;
+
+            std::cout << "Perf: " << ave_time << " ms, " << tflops << " TFlops, " << gb_per_sec
+                      << " GB/s" << std::endl;
+
+            if(tflops > best_tflops)
+            {
+                best_conv_name  = conv_name;
+                best_tflops     = tflops;
+                best_ave_time   = ave_time;
+                best_gb_per_sec = gb_per_sec;
+            }
+
+            if(do_verification)
+            {
+                in_device_buf.FromDevice(in_n_c_hi_wi_device_result.mData.data());
+
+                if(!check_out(in_n_c_hi_wi_host_result, in_n_c_hi_wi_device_result))
+                {
+                    std::cout << "Fail Info: " << conv_ptr->GetTypeString() << std::endl;
+
+                    success = false;
+                }
+                else
+                {
+                    std::cout << "Pass Info: " << conv_ptr->GetTypeString() << std::endl;
+                }
+
+                check_error(in_n_c_hi_wi_host_result, in_n_c_hi_wi_device_result);
+
+                if(do_log)
+                {
+                    std::cout << "in : ";
+                    show_data_nhwc_layout(out_n_k_ho_wo);
+                    std::cout << std::endl;
+
+                    std::cout << "wei: ";
+                    show_data_nhwc_layout(wei_k_c_y_x);
+                    std::cout << std::endl;
+
+                    std::cout << "out_host  : ";
+                    show_data_nhwc_layout(in_n_c_hi_wi_host_result);
+                    std::cout << std::endl;
+
+                    std::cout << "out_device: ";
+                    show_data_nhwc_layout(in_n_c_hi_wi_device_result);
+                    std::cout << std::endl;
+                }
+            }
+        }
+    }
+
+    std::cout << "Best Perf: " << best_ave_time << " ms, " << best_tflops << " TFlops, "
+              << best_gb_per_sec << " GB/s, " << best_conv_name << std::endl;
+    return success;
+}
+
+} // namespace profiler
+} // namespace ck

profiler/src/profile_conv_bwd_data.cpp

@@ -89,6 +89,7 @@ int profile_conv_bwd_data(int argc, char* argv[])
                                                  float,
                                                  float,
                                                  float,
+                                                 float,
                                                  ck::tensor_layout::convolution::NHWC,
                                                  ck::tensor_layout::convolution::KYXC,
                                                  ck::tensor_layout::convolution::NHWK>(
@@ -114,6 +115,7 @@ int profile_conv_bwd_data(int argc, char* argv[])
                                                  ck::half_t,
                                                  ck::half_t,
                                                  ck::half_t,
+                                                 float,
                                                  ck::tensor_layout::convolution::NHWC,
                                                  ck::tensor_layout::convolution::KYXC,
                                                  ck::tensor_layout::convolution::NHWK>(
@@ -139,6 +141,7 @@ int profile_conv_bwd_data(int argc, char* argv[])
                                                  uint16_t,
                                                  uint16_t,
                                                  uint16_t,
+                                                 float,
                                                  ck::tensor_layout::convolution::NHWC,
                                                  ck::tensor_layout::convolution::KYXC,
                                                  ck::tensor_layout::convolution::NHWK>(
@@ -164,6 +167,7 @@ int profile_conv_bwd_data(int argc, char* argv[])
                                                  int8_t,
                                                  int8_t,
                                                  int8_t,
+                                                 int32_t,
                                                  ck::tensor_layout::convolution::NHWC,
                                                  ck::tensor_layout::convolution::KYXC,
                                                  ck::tensor_layout::convolution::NHWK>(

profiler/src/profile_convnd_bwd_data.cpp

+#include <iostream>
+#include <numeric>
+#include <initializer_list>
+#include <cstdlib>
+#include <stdlib.h>
+#include <half.hpp>
+
+#include "profile_convnd_bwd_data_impl.hpp"
+
+enum ConvDataType
+{
+    F32_F32_F32,    // 0
+    F16_F16_F16,    // 1
+    BF16_BF16_BF16, // 2
+    INT8_INT8_INT8, // 3
+};
+
+enum ConvInputLayout
+{
+    NCHW, // 0
+    NHWC, // 1
+};
+
+enum ConvWeightLayout
+{
+    KCYX, // 0
+    KYXC, // 1
+};
+
+enum ConvOutputLayout
+{
+    NKHW, // 0
+    NHWK, // 1
+};
+ck::conv_util::ConvParams parse_conv_params(int num_dim_spatial, char* argv[], int arg_idx)
+{
+    // (N, K, C) + num_dim_spatial * 6 (filter, input, strides, dilations, pad left, pad right)
+    ck::conv_util::ConvParams params;
+
+    params.num_dim_spatial = num_dim_spatial;
+    params.N               = std::stoi(argv[arg_idx++]);
+    params.K               = std::stoi(argv[arg_idx++]);
+    params.C               = std::stoi(argv[arg_idx++]);
+
+    params.filter_spatial_lengths.resize(num_dim_spatial);
+    for(int i = 0; i < num_dim_spatial; ++i)
+    {
+        params.filter_spatial_lengths[i] = std::stoi(argv[arg_idx++]);
+    }
+    params.input_spatial_lengths.resize(num_dim_spatial);
+    for(int i = 0; i < num_dim_spatial; ++i)
+    {
+        params.input_spatial_lengths[i] = std::stoi(argv[arg_idx++]);
+    }
+    params.conv_filter_strides.resize(num_dim_spatial);
+    for(int i = 0; i < num_dim_spatial; ++i)
+    {
+        params.conv_filter_strides[i] = std::stoi(argv[arg_idx++]);
+    }
+    params.conv_filter_dilations.resize(num_dim_spatial);
+    for(int i = 0; i < num_dim_spatial; ++i)
+    {
+        params.conv_filter_dilations[i] = std::stoi(argv[arg_idx++]);
+    }
+    params.input_left_pads.resize(num_dim_spatial);
+    for(int i = 0; i < num_dim_spatial; ++i)
+    {
+        params.input_left_pads[i] = std::stoi(argv[arg_idx++]);
+    }
+    params.input_right_pads.resize(num_dim_spatial);
+    for(int i = 0; i < num_dim_spatial; ++i)
+    {
+        params.input_right_pads[i] = std::stoi(argv[arg_idx++]);
+    }
+
+    return params;
+}
+
+int profile_convnd_bwd_data(int argc, char* argv[], int num_dim_spatial)
+{
+    const int preParams = 10;
+    int conv_args       = 3 + num_dim_spatial * 6;
+    int cmdline_nargs   = conv_args + preParams;
+    if(cmdline_nargs != argc)
+    {
+        printf("arg1: tensor operation (conv[1|2|3]d_bwd_data: BackwardConvolution)\n");
+        printf("arg2: data type (0: fp32; 1: fp16)\n");
+        printf("arg3: input tensor layout (0: NCHW; 1: NHWC)\n");
+        printf("arg4: weight tensor layout (0: KCYX; 1: KYXC)\n");
+        printf("arg5: output tensor layout (0: NKHW; 1: NHWK)\n");
+        printf("arg6: verification (0: no; 1: yes)\n");
+        printf("arg7: initialization (0: no init; 1: integer value; 2: decimal value)\n");
+        printf("arg8: print tensor value (0: no; 1: yes)\n");
+        printf("arg9: run kernel # of times (>1)\n");
+        printf("arg10 to 24: N, K, C, Y, X, Hi, Wi, Sy, Sx, Dy, Dx, LeftPy, LeftPx, RightPy, "
+               "RightPx\n");
+        return 1;
+    }
+
+    const int data_type        = static_cast<ConvDataType>(std::stoi(argv[2]));
+    const int in_layout        = static_cast<ConvInputLayout>(std::stoi(argv[3]));
+    const int wei_layout       = static_cast<ConvWeightLayout>(std::stoi(argv[4]));
+    const int out_layout       = static_cast<ConvOutputLayout>(std::stoi(argv[5]));
+    const bool do_verification = std::stoi(argv[6]);
+    const int init_method      = std::stoi(argv[7]);
+    const bool do_log          = std::stoi(argv[8]);
+    const int nrepeat          = std::stoi(argv[9]);
+
+    ck::conv_util::ConvParams params = parse_conv_params(num_dim_spatial, argv, preParams);
+
+    auto Run = [&](auto input_type, auto wei_type, auto out_type, auto acc_type) {
+        using InDataType  = decltype(input_type);
+        using WeiDataType = decltype(wei_type);
+        using OutDataType = decltype(out_type);
+        using AccDataType = decltype(acc_type);
+
+        switch(num_dim_spatial)
+        {
+        case 1:
+            ck::profiler::profile_convnd_bwd_data_impl<1,
+                                                       InDataType,
+                                                       WeiDataType,
+                                                       OutDataType,
+                                                       AccDataType,
+                                                       ck::tensor_layout::convolution::NWC,
+                                                       ck::tensor_layout::convolution::KXC,
+                                                       ck::tensor_layout::convolution::NWK>(
+                do_verification,
+                init_method,
+                do_log,
+                nrepeat,
+                params.N,
+                params.K,
+                params.C,
+                params.input_spatial_lengths,
+                params.filter_spatial_lengths,
+                params.GetOutputSpatialLengths(),
+                params.conv_filter_strides,
+                params.conv_filter_dilations,
+                params.input_left_pads,
+                params.input_right_pads);
+            break;
+
+        case 2:
+            ck::profiler::profile_convnd_bwd_data_impl<2,
+                                                       InDataType,
+                                                       WeiDataType,
+                                                       OutDataType,
+                                                       AccDataType,
+                                                       ck::tensor_layout::convolution::NHWC,
+                                                       ck::tensor_layout::convolution::KYXC,
+                                                       ck::tensor_layout::convolution::NHWK>(
+                do_verification,
+                init_method,
+                do_log,
+                nrepeat,
+                params.N,
+                params.K,
+                params.C,
+                params.input_spatial_lengths,
+                params.filter_spatial_lengths,
+                params.GetOutputSpatialLengths(),
+                params.conv_filter_strides,
+                params.conv_filter_dilations,
+                params.input_left_pads,
+                params.input_right_pads);
+            break;
+
+        case 3:
+            ck::profiler::profile_convnd_bwd_data_impl<3,
+                                                       InDataType,
+                                                       WeiDataType,
+                                                       OutDataType,
+                                                       AccDataType,
+                                                       ck::tensor_layout::convolution::NDHWC,
+                                                       ck::tensor_layout::convolution::KZYXC,
+                                                       ck::tensor_layout::convolution::NDHWK>(
+                do_verification,
+                init_method,
+                do_log,
+                nrepeat,
+                params.N,
+                params.K,
+                params.C,
+                params.input_spatial_lengths,
+                params.filter_spatial_lengths,
+                params.GetOutputSpatialLengths(),
+                params.conv_filter_strides,
+                params.conv_filter_dilations,
+                params.input_left_pads,
+                params.input_right_pads);
+            break;
+
+        default: break;
+        }
+    };
+    if(data_type == ConvDataType::F32_F32_F32 && in_layout == ConvInputLayout::NHWC &&
+       wei_layout == ConvWeightLayout::KYXC && out_layout == ConvOutputLayout::NHWK)
+    {
+        Run(float{}, float{}, float{}, float{});
+    }
+    else if(data_type == ConvDataType::F16_F16_F16 && in_layout == ConvInputLayout::NHWC &&
+            wei_layout == ConvWeightLayout::KYXC && out_layout == ConvOutputLayout::NHWK)
+    {
+        Run(ck::half_t{}, ck::half_t{}, ck::half_t{}, float{});
+    }
+    else if(data_type == ConvDataType::BF16_BF16_BF16 && in_layout == ConvInputLayout::NHWC &&
+            wei_layout == ConvWeightLayout::KYXC && out_layout == ConvOutputLayout::NHWK)
+    {
+        Run(ck::bhalf_t{}, ck::bhalf_t{}, ck::bhalf_t{}, float{});
+    }
+    else if(data_type == ConvDataType::INT8_INT8_INT8 && in_layout == ConvInputLayout::NHWC &&
+            wei_layout == ConvWeightLayout::KYXC && out_layout == ConvOutputLayout::NHWK)
+    {
+        Run(int8_t{}, int8_t{}, int8_t{}, int32_t{});
+    }
+    else
+    {
+        std::cout << "wrong! this Conv data_type & layout is not implemented" << std::endl;
+        return 1;
+    }
+
+    return 0;
+}

profiler/src/profiler.cpp

@@ -15,7 +15,7 @@ int profile_conv_fwd(int, char*[]);
 int profile_conv_fwd_bias_relu(int, char*[]);
 int profile_conv_fwd_bias_relu_add(int, char*[]);
 int profile_conv_fwd_bias_relu_atomic_add(int, char*[]);
-int profile_conv_bwd_data(int, char*[]);
+int profile_convnd_bwd_data(int, char*[], int);
 int profile_reduce(int, char*[]);
 
 int main(int argc, char* argv[])
@@ -64,9 +64,17 @@ int main(int argc, char* argv[])
     {
         return profile_conv_fwd_bias_relu_atomic_add(argc, argv);
     }
-    else if(strcmp(argv[1], "conv_bwd") == 0)
+    else if(strcmp(argv[1], "conv1d_bwd_data") == 0)
     {
-        return profile_conv_bwd_data(argc, argv);
+        return profile_convnd_bwd_data(argc, argv, 1);
+    }
+    else if(strcmp(argv[1], "conv2d_bwd_data") == 0)
+    {
+        return profile_convnd_bwd_data(argc, argv, 2);
+    }
+    else if(strcmp(argv[1], "conv3d_bwd_data") == 0)
+    {
+        return profile_convnd_bwd_data(argc, argv, 3);
     }
     else if(strcmp(argv[1], "reduce") == 0)
     {
@@ -85,8 +93,11 @@ int main(int argc, char* argv[])
                "                        conv_fwd_bias_relu: ForwardConvolution+Bias+ReLU\n"
                "                        conv_fwd_bias_relu_add: ForwardConvolution+Bias+ReLU+Add\n"
                "                        conv_fwd_bias_relu_atomic_add: ForwardConvolution+Bias+ReLU+AtomicAdd\n"
-               "                        conv_bwd: BackwardConvolution\n"
-               "                        reduce: Reduce\n");
+               "                        conv1d_bwd_data: BackwardConvolution data 1 dim\n"
+               "                        conv2d_bwd_data: BackwardConvolution data 2 dim\n"
+               "                        conv3d_bwd_data: BackwardConvolution data 3 dim\n"
+               "                        grouped_gemm: Grouped Gemm\n"
+               "                        reduce: REDUCE\n");
         // clang-format on
 
         return 0;

test/CMakeLists.txt

@@ -42,6 +42,5 @@ add_subdirectory(gemm_reduce)
 add_subdirectory(batched_gemm)
 add_subdirectory(grouped_gemm)
 add_subdirectory(convnd_fwd)
-add_subdirectory(conv2d_bwd_data)
 add_subdirectory(reduce)
 add_subdirectory(cpu_ukernel)

test/conv2d_bwd_data/conv2d_bwd_data.cpp

@@ -121,15 +121,17 @@ int main(int argc, char* argv[])
         exit(1);
     }
 
-    auto Run = [&](auto input_type, auto wei_type, auto out_type) {
+    auto Run = [&](auto input_type, auto wei_type, auto out_type, auto acc_type) {
         using InDataType  = decltype(input_type);
         using WeiDataType = decltype(wei_type);
         using OutDataType = decltype(out_type);
+        using AccDataType = decltype(acc_type);
 
         using ReferenceConvBwdInstance =
             ck::tensor_operation::host::ReferenceConvBwdData<InDataType,
                                                              WeiDataType,
                                                              OutDataType,
+                                                             AccDataType,
                                                              InElementOp,
                                                              WeiElementOp,
                                                              OutElementOp>;
@@ -293,33 +295,33 @@ int main(int argc, char* argv[])
         if(success)
         {
             std::cout << "test conv2d bwd : Pass" << std::endl;
+            return 0;
         }
         else
         {
             std::cout << "test conv2d bwd: Fail " << std::endl;
+            return -1;
         }
     };
 
     if(data_type == 0)
     {
-        Run(F32(), F32(), F32());
+        return Run(F32(), F32(), F32(), F32());
     }
     else if(data_type == 1)
     {
-        Run(F16(), F16(), F16());
+        return Run(F16(), F16(), F16(), F32());
     }
     else if(data_type == 2)
     {
-        Run(BF16(), BF16(), BF16());
+        return Run(BF16(), BF16(), BF16(), F32());
     }
     else if(data_type == 3)
     {
-        Run(INT8(), INT8(), INT8());
+        return Run(INT8(), INT8(), INT8(), int());
     }
     else
     {
         return 1;
     }
-
-    return 0;
 }

test/convnd_bwd_data/CMakeLists.txt

+include_directories(BEFORE
+    ${PROJECT_SOURCE_DIR}/profiler/include
+    ${PROJECT_SOURCE_DIR}/external/include/half
+)
+
+add_test_executable(test_convnd_bwd_data convnd_bwd_data.cpp)
+target_link_libraries(test_convnd_bwd_data PRIVATE host_tensor)
+target_link_libraries(test_convnd_bwd_data PRIVATE device_convnd_bwd_data_instance)

test/convnd_bwd_data/convnd_bwd_data.cpp

+#include <iostream>
+#include <numeric>
+#include <initializer_list>
+#include <cstdlib>
+#include <stdlib.h>
+#include <half.hpp>
+#include <vector>
+
+#include "profile_convnd_bwd_data_impl.hpp"
+
+int main()
+{
+    bool pass = true;
+    // check 1d
+    std::vector<ck::conv_util::ConvParams> params;
+    params.push_back({1, 128, 128, 256, {1}, {14}, {2}, {1}, {0}, {0}});
+    params.push_back({1, 128, 128, 256, {3}, {28}, {1}, {1}, {1}, {1}});
+    params.push_back({1, 128, 128, 256, {1}, {3}, {1}, {1}, {0}, {0}});
+
+    for(auto& param : params)
+    {
+        pass &= ck::profiler::profile_convnd_bwd_data_impl<1,
+                                                           float,
+                                                           float,
+                                                           float,
+                                                           float,
+                                                           ck::tensor_layout::convolution::NWC,
+                                                           ck::tensor_layout::convolution::KXC,
+                                                           ck::tensor_layout::convolution::NWK>(
+            1, // do_verification,
+            1, // init_method,
+            0, // do_log,
+            1, // nrepeat,
+            param.N,
+            param.K,
+            param.C,
+            param.input_spatial_lengths,
+            param.filter_spatial_lengths,
+            param.GetOutputSpatialLengths(),
+            param.conv_filter_strides,
+            param.conv_filter_dilations,
+            param.input_left_pads,
+            param.input_right_pads);
+
+        pass &= ck::profiler::profile_convnd_bwd_data_impl<1,
+                                                           ck::half_t,
+                                                           ck::half_t,
+                                                           ck::half_t,
+                                                           float,
+                                                           ck::tensor_layout::convolution::NWC,
+                                                           ck::tensor_layout::convolution::KXC,
+                                                           ck::tensor_layout::convolution::NWK>(
+            1, // do_verification,
+            1, // init_method,
+            0, // do_log,
+            1, // nrepeat,
+            param.N,
+            param.K,
+            param.C,
+            param.input_spatial_lengths,
+            param.filter_spatial_lengths,
+            param.GetOutputSpatialLengths(),
+            param.conv_filter_strides,
+            param.conv_filter_dilations,
+            param.input_left_pads,
+            param.input_right_pads);
+
+        pass &= ck::profiler::profile_convnd_bwd_data_impl<1,
+                                                           ck::bhalf_t,
+                                                           ck::bhalf_t,
+                                                           ck::bhalf_t,
+                                                           float,
+                                                           ck::tensor_layout::convolution::NWC,
+                                                           ck::tensor_layout::convolution::KXC,
+                                                           ck::tensor_layout::convolution::NWK>(
+            1, // do_verification,
+            1, // init_method,
+            0, // do_log,
+            1, // nrepeat,
+            param.N,
+            param.K,
+            param.C,
+            param.input_spatial_lengths,
+            param.filter_spatial_lengths,
+            param.GetOutputSpatialLengths(),
+            param.conv_filter_strides,
+            param.conv_filter_dilations,
+            param.input_left_pads,
+            param.input_right_pads);
+
+        pass &= ck::profiler::profile_convnd_bwd_data_impl<1,
+                                                           int8_t,
+                                                           int8_t,
+                                                           int8_t,
+                                                           int,
+                                                           ck::tensor_layout::convolution::NWC,
+                                                           ck::tensor_layout::convolution::KXC,
+                                                           ck::tensor_layout::convolution::NWK>(
+            1, // do_verification,
+            1, // init_method,
+            0, // do_log,
+            1, // nrepeat,
+            param.N,
+            param.K,
+            param.C,
+            param.input_spatial_lengths,
+            param.filter_spatial_lengths,
+            param.GetOutputSpatialLengths(),
+            param.conv_filter_strides,
+            param.conv_filter_dilations,
+            param.input_left_pads,
+            param.input_right_pads);
+    }
+
+    // check 2d
+    params.clear();
+    params.push_back({2, 128, 128, 256, {1, 1}, {7, 7}, {2, 2}, {1, 1}, {0, 0}, {0, 0}});
+    params.push_back({2, 128, 128, 256, {3, 3}, {14, 14}, {1, 1}, {1, 1}, {1, 1}, {1, 1}});
+    params.push_back({2, 128, 128, 256, {1, 1}, {3, 3}, {1, 1}, {1, 1}, {0, 0}, {0, 0}});
+
+    for(auto& param : params)
+    {
+        pass &= ck::profiler::profile_convnd_bwd_data_impl<2,
+                                                           float,
+                                                           float,
+                                                           float,
+                                                           float,
+                                                           ck::tensor_layout::convolution::NHWC,
+                                                           ck::tensor_layout::convolution::KYXC,
+                                                           ck::tensor_layout::convolution::NHWK>(
+            1, // do_verification,
+            1, // init_method,
+            0, // do_log,
+            1, // nrepeat,
+            param.N,
+            param.K,
+            param.C,
+            param.input_spatial_lengths,
+            param.filter_spatial_lengths,
+            param.GetOutputSpatialLengths(),
+            param.conv_filter_strides,
+            param.conv_filter_dilations,
+            param.input_left_pads,
+            param.input_right_pads);
+
+        pass &= ck::profiler::profile_convnd_bwd_data_impl<2,
+                                                           ck::half_t,
+                                                           ck::half_t,
+                                                           ck::half_t,
+                                                           float,
+                                                           ck::tensor_layout::convolution::NHWC,
+                                                           ck::tensor_layout::convolution::KYXC,
+                                                           ck::tensor_layout::convolution::NHWK>(
+            1, // do_verification,
+            1, // init_method,
+            0, // do_log,
+            1, // nrepeat,
+            param.N,
+            param.K,
+            param.C,
+            param.input_spatial_lengths,
+            param.filter_spatial_lengths,
+            param.GetOutputSpatialLengths(),
+            param.conv_filter_strides,
+            param.conv_filter_dilations,
+            param.input_left_pads,
+            param.input_right_pads);
+
+        pass &= ck::profiler::profile_convnd_bwd_data_impl<2,
+                                                           ck::bhalf_t,
+                                                           ck::bhalf_t,
+                                                           ck::bhalf_t,
+                                                           float,
+                                                           ck::tensor_layout::convolution::NHWC,
+                                                           ck::tensor_layout::convolution::KYXC,
+                                                           ck::tensor_layout::convolution::NHWK>(
+            1, // do_verification,
+            1, // init_method,
+            0, // do_log,
+            1, // nrepeat,
+            param.N,
+            param.K,
+            param.C,
+            param.input_spatial_lengths,
+            param.filter_spatial_lengths,
+            param.GetOutputSpatialLengths(),
+            param.conv_filter_strides,
+            param.conv_filter_dilations,
+            param.input_left_pads,
+            param.input_right_pads);
+
+        pass &= ck::profiler::profile_convnd_bwd_data_impl<2,
+                                                           int8_t,
+                                                           int8_t,
+                                                           int8_t,
+                                                           int,
+                                                           ck::tensor_layout::convolution::NHWC,
+                                                           ck::tensor_layout::convolution::KYXC,
+                                                           ck::tensor_layout::convolution::NHWK>(
+            1, // do_verification,
+            1, // init_method,
+            0, // do_log,
+            1, // nrepeat,
+            param.N,
+            param.K,
+            param.C,
+            param.input_spatial_lengths,
+            param.filter_spatial_lengths,
+            param.GetOutputSpatialLengths(),
+            param.conv_filter_strides,
+            param.conv_filter_dilations,
+            param.input_left_pads,
+            param.input_right_pads);
+    }
+
+    // check 3d
+    params.clear();
+    params.push_back(
+        {3, 128, 128, 256, {1, 1, 1}, {7, 7, 7}, {2, 2, 2}, {1, 1, 1}, {0, 0, 0}, {0, 0, 0}});
+    params.push_back(
+        {3, 128, 128, 256, {3, 3, 3}, {14, 14, 14}, {1, 1, 1}, {1, 1, 1}, {1, 1, 1}, {1, 1, 1}});
+    params.push_back(
+        {3, 128, 128, 256, {1, 1, 1}, {3, 3, 3}, {1, 1, 1}, {1, 1, 1}, {0, 0, 0}, {0, 0, 0}});
+
+    for(auto& param : params)
+    {
+        pass &= ck::profiler::profile_convnd_bwd_data_impl<3,
+                                                           float,
+                                                           float,
+                                                           float,
+                                                           float,
+                                                           ck::tensor_layout::convolution::NDHWC,
+                                                           ck::tensor_layout::convolution::KZYXC,
+                                                           ck::tensor_layout::convolution::NDHWK>(
+            1, // do_verification,
+            1, // init_method,
+            0, // do_log,
+            1, // nrepeat,
+            param.N,
+            param.K,
+            param.C,
+            param.input_spatial_lengths,
+            param.filter_spatial_lengths,
+            param.GetOutputSpatialLengths(),
+            param.conv_filter_strides,
+            param.conv_filter_dilations,
+            param.input_left_pads,
+            param.input_right_pads);
+
+        pass &= ck::profiler::profile_convnd_bwd_data_impl<3,
+                                                           ck::half_t,
+                                                           ck::half_t,
+                                                           ck::half_t,
+                                                           float,
+                                                           ck::tensor_layout::convolution::NDHWC,
+                                                           ck::tensor_layout::convolution::KZYXC,
+                                                           ck::tensor_layout::convolution::NDHWK>(
+            1, // do_verification,
+            1, // init_method,
+            0, // do_log,
+            1, // nrepeat,
+            param.N,
+            param.K,
+            param.C,
+            param.input_spatial_lengths,
+            param.filter_spatial_lengths,
+            param.GetOutputSpatialLengths(),
+            param.conv_filter_strides,
+            param.conv_filter_dilations,
+            param.input_left_pads,
+            param.input_right_pads);
+
+        pass &= ck::profiler::profile_convnd_bwd_data_impl<3,
+                                                           ck::bhalf_t,
+                                                           ck::bhalf_t,
+                                                           ck::bhalf_t,
+                                                           float,
+                                                           ck::tensor_layout::convolution::NDHWC,
+                                                           ck::tensor_layout::convolution::KZYXC,
+                                                           ck::tensor_layout::convolution::NDHWK>(
+            1, // do_verification,
+            1, // init_method,
+            0, // do_log,
+            1, // nrepeat,
+            param.N,
+            param.K,
+            param.C,
+            param.input_spatial_lengths,
+            param.filter_spatial_lengths,
+            param.GetOutputSpatialLengths(),
+            param.conv_filter_strides,
+            param.conv_filter_dilations,
+            param.input_left_pads,
+            param.input_right_pads);
+
+        pass &= ck::profiler::profile_convnd_bwd_data_impl<3,
+                                                           int8_t,
+                                                           int8_t,
+                                                           int8_t,
+                                                           int,
+                                                           ck::tensor_layout::convolution::NDHWC,
+                                                           ck::tensor_layout::convolution::KZYXC,
+                                                           ck::tensor_layout::convolution::NDHWK>(
+            1, // do_verification,
+            1, // init_method,
+            0, // do_log,
+            1, // nrepeat,
+            param.N,
+            param.K,
+            param.C,
+            param.input_spatial_lengths,
+            param.filter_spatial_lengths,
+            param.GetOutputSpatialLengths(),
+            param.conv_filter_strides,
+            param.conv_filter_dilations,
+            param.input_left_pads,
+            param.input_right_pads);
+    }
+
+    if(pass)
+    {
+        std::cout << "test convnd bwd : Pass" << std::endl;
+        return 0;
+    }
+    else
+    {
+        std::cout << "test convnd bwd: Fail " << std::endl;
+        return -1;
+    }
+}