Add grouped conv bwd weight dl instances and new layout (#897)

* Add grouped conv bwd weight dl instances and new layout

* Add M and N padding

* Remove todo comment

* Enable grouped conv fwd dl k,c=1 generic instance

* Comment fixes

library/src/tensor_operation_instance/gpu/grouped_conv2d_bwd_weight/device_grouped_conv2d_bwd_weight_dl_gnhwc_gkyxc_gnhwk_f32_instance.cpp

+// SPDX-License-Identifier: MIT
+// Copyright (c) 2018-2023, Advanced Micro Devices, Inc. All rights reserved.
+
+#include "ck/library/tensor_operation_instance/add_device_operation_instance.hpp"
+#include "ck/library/tensor_operation_instance/gpu/grouped_conv_bwd_weight/device_grouped_conv_bwd_weight_dl_instance.hpp"
+
+namespace ck {
+namespace tensor_operation {
+namespace device {
+namespace instance {
+
+// Compilation parameters for in[g, n, hi, wi, c] * wei[g, k, y, x, c] = out[g, n, ho, wo, k]
+void add_device_grouped_conv2d_bwd_weight_dl_gnhwc_gkyxc_gnhwk_f32_instances(
+    std::vector<std::unique_ptr<DeviceGroupedConvBwdWeight<2,
+                                                           GNHWC,
+                                                           GKYXC,
+                                                           GNHWK,
+                                                           F32,
+                                                           F32,
+                                                           F32,
+                                                           PassThrough,
+                                                           PassThrough,
+                                                           PassThrough>>>& instances)
+{
+    // 1. Default
+    add_device_operation_instances(
+        instances,
+        device_grouped_conv_bwd_weight_dl_f32_instances<2,
+                                                        GNHWC,
+                                                        GKYXC,
+                                                        GNHWK,
+                                                        ConvBwdWeightDefault>{});
+    // 2. Filter1x1Stride1Pad0
+    add_device_operation_instances(
+        instances,
+        device_grouped_conv_bwd_weight_dl_f32_instances<2,
+                                                        GNHWC,
+                                                        GKYXC,
+                                                        GNHWK,
+                                                        ConvBwdWeightFilter1x1Stride1Pad0>{});
+}
+
+} // namespace instance
+} // namespace device
+} // namespace tensor_operation
+} // namespace ck

library/src/tensor_operation_instance/gpu/grouped_conv2d_bwd_weight/device_grouped_conv2d_bwd_weight_dl_nhwgc_gkyxc_nhwgk_bf16_instance.cpp

+// SPDX-License-Identifier: MIT
+// Copyright (c) 2018-2023, Advanced Micro Devices, Inc. All rights reserved.
+
+#include "ck/library/tensor_operation_instance/add_device_operation_instance.hpp"
+#include "ck/library/tensor_operation_instance/gpu/grouped_conv_bwd_weight/device_grouped_conv_bwd_weight_dl_instance.hpp"
+
+namespace ck {
+namespace tensor_operation {
+namespace device {
+namespace instance {
+
+// Compilation parameters for in[n, hi, wi, g, c] * wei[g, k, y, x, c] = out[n, ho, wo, g, k]
+void add_device_grouped_conv2d_bwd_weight_dl_nhwgc_gkyxc_nhwgk_bf16_f32_bf16_instances(
+    std::vector<std::unique_ptr<DeviceGroupedConvBwdWeight<2,
+                                                           NHWGC,
+                                                           GKYXC,
+                                                           NHWGK,
+                                                           BF16,
+                                                           F32,
+                                                           BF16,
+                                                           PassThrough,
+                                                           PassThrough,
+                                                           PassThrough>>>& instances)
+{
+    // 1. Default
+    add_device_operation_instances(
+        instances,
+        device_grouped_conv_bwd_weight_dl_bf16_instances<2,
+                                                         NHWGC,
+                                                         GKYXC,
+                                                         NHWGK,
+                                                         ConvBwdWeightDefault>{});
+    // 2. Filter1x1Stride1Pad0
+    add_device_operation_instances(
+        instances,
+        device_grouped_conv_bwd_weight_dl_bf16_instances<2,
+                                                         NHWGC,
+                                                         GKYXC,
+                                                         NHWGK,
+                                                         ConvBwdWeightFilter1x1Stride1Pad0>{});
+}
+
+} // namespace instance
+} // namespace device
+} // namespace tensor_operation
+} // namespace ck

library/src/tensor_operation_instance/gpu/grouped_conv2d_bwd_weight/device_grouped_conv2d_bwd_weight_dl_nhwgc_gkyxc_nhwgk_f16_instance.cpp

+// SPDX-License-Identifier: MIT
+// Copyright (c) 2018-2023, Advanced Micro Devices, Inc. All rights reserved.
+
+#include "ck/library/tensor_operation_instance/add_device_operation_instance.hpp"
+#include "ck/library/tensor_operation_instance/gpu/grouped_conv_bwd_weight/device_grouped_conv_bwd_weight_dl_instance.hpp"
+
+namespace ck {
+namespace tensor_operation {
+namespace device {
+namespace instance {
+
+// Compilation parameters for in[n, hi, wi, g, c] * wei[g, k, y, x, c] = out[n, ho, wo, g, k]
+void add_device_grouped_conv2d_bwd_weight_dl_nhwgc_gkyxc_nhwgk_f16_instances(
+    std::vector<std::unique_ptr<DeviceGroupedConvBwdWeight<2,
+                                                           NHWGC,
+                                                           GKYXC,
+                                                           NHWGK,
+                                                           F16,
+                                                           F16,
+                                                           F16,
+                                                           PassThrough,
+                                                           PassThrough,
+                                                           PassThrough>>>& instances)
+{
+    // 1. Default
+    add_device_operation_instances(
+        instances,
+        device_grouped_conv_bwd_weight_dl_f16_instances<2,
+                                                        NHWGC,
+                                                        GKYXC,
+                                                        NHWGK,
+                                                        ConvBwdWeightDefault>{});
+    // 2. Filter1x1Stride1Pad0
+    add_device_operation_instances(
+        instances,
+        device_grouped_conv_bwd_weight_dl_f16_instances<2,
+                                                        NHWGC,
+                                                        GKYXC,
+                                                        NHWGK,
+                                                        ConvBwdWeightFilter1x1Stride1Pad0>{});
+}
+
+} // namespace instance
+} // namespace device
+} // namespace tensor_operation
+} // namespace ck

library/src/tensor_operation_instance/gpu/grouped_conv2d_bwd_weight/device_grouped_conv2d_bwd_weight_dl_nhwgc_gkyxc_nhwgk_f32_instance.cpp

+// SPDX-License-Identifier: MIT
+// Copyright (c) 2018-2023, Advanced Micro Devices, Inc. All rights reserved.
+
+#include "ck/library/tensor_operation_instance/add_device_operation_instance.hpp"
+#include "ck/library/tensor_operation_instance/gpu/grouped_conv_bwd_weight/device_grouped_conv_bwd_weight_dl_instance.hpp"
+
+namespace ck {
+namespace tensor_operation {
+namespace device {
+namespace instance {
+
+// Compilation parameters for in[n, hi, wi, g, c] * wei[g, k, y, x, c] = out[n, ho, wo, g, k]
+void add_device_grouped_conv2d_bwd_weight_dl_nhwgc_gkyxc_nhwgk_f32_instances(
+    std::vector<std::unique_ptr<DeviceGroupedConvBwdWeight<2,
+                                                           NHWGC,
+                                                           GKYXC,
+                                                           NHWGK,
+                                                           F32,
+                                                           F32,
+                                                           F32,
+                                                           PassThrough,
+                                                           PassThrough,
+                                                           PassThrough>>>& instances)
+{
+    // 1. Default
+    add_device_operation_instances(
+        instances,
+        device_grouped_conv_bwd_weight_dl_f32_instances<2,
+                                                        NHWGC,
+                                                        GKYXC,
+                                                        NHWGK,
+                                                        ConvBwdWeightDefault>{});
+    // 2. Filter1x1Stride1Pad0
+    add_device_operation_instances(
+        instances,
+        device_grouped_conv_bwd_weight_dl_f32_instances<2,
+                                                        NHWGC,
+                                                        GKYXC,
+                                                        NHWGK,
+                                                        ConvBwdWeightFilter1x1Stride1Pad0>{});
+}
+
+} // namespace instance
+} // namespace device
+} // namespace tensor_operation
+} // namespace ck

library/src/tensor_operation_instance/gpu/grouped_conv2d_bwd_weight/device_grouped_conv2d_bwd_weight_xdl_gnhwc_gkyxc_gnhwk_bf16_instance.cpp

@@ -2,7 +2,6 @@
 // Copyright (c) 2018-2023, Advanced Micro Devices, Inc. All rights reserved.
 
 #include "ck/library/tensor_operation_instance/add_device_operation_instance.hpp"
-
 #include "ck/library/tensor_operation_instance/gpu/grouped_conv_bwd_weight/device_grouped_conv_bwd_weight_xdl_instance.hpp"
 
 namespace ck {

library/src/tensor_operation_instance/gpu/grouped_conv2d_bwd_weight/device_grouped_conv2d_bwd_weight_xdl_gnhwc_gkyxc_gnhwk_f16_instance.cpp

@@ -2,7 +2,6 @@
 // Copyright (c) 2018-2023, Advanced Micro Devices, Inc. All rights reserved.
 
 #include "ck/library/tensor_operation_instance/add_device_operation_instance.hpp"
-
 #include "ck/library/tensor_operation_instance/gpu/grouped_conv_bwd_weight/device_grouped_conv_bwd_weight_xdl_instance.hpp"
 
 namespace ck {

library/src/tensor_operation_instance/gpu/grouped_conv2d_bwd_weight/device_grouped_conv2d_bwd_weight_xdl_gnhwc_gkyxc_gnhwk_f32_instance.cpp

@@ -2,7 +2,6 @@
 // Copyright (c) 2018-2023, Advanced Micro Devices, Inc. All rights reserved.
 
 #include "ck/library/tensor_operation_instance/add_device_operation_instance.hpp"
-
 #include "ck/library/tensor_operation_instance/gpu/grouped_conv_bwd_weight/device_grouped_conv_bwd_weight_xdl_instance.hpp"
 
 namespace ck {

library/src/tensor_operation_instance/gpu/grouped_conv2d_bwd_weight/device_grouped_conv2d_bwd_weight_xdl_nhwgc_gkyxc_nhwgk_bf16_instance.cpp

@@ -2,7 +2,6 @@
 // Copyright (c) 2018-2023, Advanced Micro Devices, Inc. All rights reserved.
 
 #include "ck/library/tensor_operation_instance/add_device_operation_instance.hpp"
-
 #include "ck/library/tensor_operation_instance/gpu/grouped_conv_bwd_weight/device_grouped_conv_bwd_weight_xdl_instance.hpp"
 
 namespace ck {

library/src/tensor_operation_instance/gpu/grouped_conv2d_bwd_weight/device_grouped_conv2d_bwd_weight_xdl_nhwgc_gkyxc_nhwgk_f16_instance.cpp

@@ -2,7 +2,6 @@
 // Copyright (c) 2018-2023, Advanced Micro Devices, Inc. All rights reserved.
 
 #include "ck/library/tensor_operation_instance/add_device_operation_instance.hpp"
-
 #include "ck/library/tensor_operation_instance/gpu/grouped_conv_bwd_weight/device_grouped_conv_bwd_weight_xdl_instance.hpp"
 
 namespace ck {

library/src/tensor_operation_instance/gpu/grouped_conv2d_bwd_weight/device_grouped_conv2d_bwd_weight_xdl_nhwgc_gkyxc_nhwgk_f32_instance.cpp

@@ -2,7 +2,6 @@
 // Copyright (c) 2018-2023, Advanced Micro Devices, Inc. All rights reserved.
 
 #include "ck/library/tensor_operation_instance/add_device_operation_instance.hpp"
-
 #include "ck/library/tensor_operation_instance/gpu/grouped_conv_bwd_weight/device_grouped_conv_bwd_weight_xdl_instance.hpp"
 
 namespace ck {

library/src/tensor_operation_instance/gpu/grouped_conv3d_bwd_weight/CMakeLists.txt

-add_instance_library(device_grouped_conv3d_bwd_weight_instance
+set(GROUPED_CONV3D_BWD_WEIGHT
     device_grouped_conv3d_bwd_weight_xdl_gndhwc_gkzyxc_gndhwk_f16_instance.cpp
     device_grouped_conv3d_bwd_weight_xdl_gndhwc_gkzyxc_gndhwk_f32_instance.cpp
     device_grouped_conv3d_bwd_weight_xdl_gndhwc_gkzyxc_gndhwk_bf16_instance.cpp
     device_grouped_conv3d_bwd_weight_xdl_ndhwgc_gkzyxc_ndhwgk_f16_instance.cpp
     device_grouped_conv3d_bwd_weight_xdl_ndhwgc_gkzyxc_ndhwgk_f32_instance.cpp
-    device_grouped_conv3d_bwd_weight_xdl_ndhwgc_gkzyxc_ndhwgk_bf16_instance.cpp
-)
+    device_grouped_conv3d_bwd_weight_xdl_ndhwgc_gkzyxc_ndhwgk_bf16_instance.cpp)
+
+if(DL_KERNELS)
+    list(APPEND GROUPED_CONV3D_BWD_WEIGHT
+        device_grouped_conv3d_bwd_weight_dl_gndhwc_gkzyxc_gndhwk_f16_instance.cpp
+        device_grouped_conv3d_bwd_weight_dl_gndhwc_gkzyxc_gndhwk_f32_instance.cpp
+        device_grouped_conv3d_bwd_weight_dl_gndhwc_gkzyxc_gndhwk_bf16_instance.cpp
+        device_grouped_conv3d_bwd_weight_dl_ndhwgc_gkzyxc_ndhwgk_f16_instance.cpp
+        device_grouped_conv3d_bwd_weight_dl_ndhwgc_gkzyxc_ndhwgk_f32_instance.cpp
+        device_grouped_conv3d_bwd_weight_dl_ndhwgc_gkzyxc_ndhwgk_bf16_instance.cpp)
+endif()
+
+add_instance_library(device_grouped_conv3d_bwd_weight_instance ${GROUPED_CONV3D_BWD_WEIGHT})

library/src/tensor_operation_instance/gpu/grouped_conv3d_bwd_weight/device_grouped_conv3d_bwd_weight_dl_gndhwc_gkzyxc_gndhwk_bf16_instance.cpp

+// SPDX-License-Identifier: MIT
+// Copyright (c) 2018-2023, Advanced Micro Devices, Inc. All rights reserved.
+
+#include "ck/library/tensor_operation_instance/add_device_operation_instance.hpp"
+#include "ck/library/tensor_operation_instance/gpu/grouped_conv_bwd_weight/device_grouped_conv_bwd_weight_dl_instance.hpp"
+
+namespace ck {
+namespace tensor_operation {
+namespace device {
+namespace instance {
+
+void add_device_grouped_conv3d_bwd_weight_dl_gndhwc_gkzyxc_gndhwk_bf16_f32_bf16_instances(
+    std::vector<std::unique_ptr<DeviceGroupedConvBwdWeight<3,
+                                                           GNDHWC,
+                                                           GKZYXC,
+                                                           GNDHWK,
+                                                           BF16,
+                                                           F32,
+                                                           BF16,
+                                                           PassThrough,
+                                                           PassThrough,
+                                                           PassThrough>>>& instances)
+{
+    // 1. Default
+    add_device_operation_instances(
+        instances,
+        device_grouped_conv_bwd_weight_dl_bf16_instances<3,
+                                                         GNDHWC,
+                                                         GKZYXC,
+                                                         GNDHWK,
+                                                         ConvBwdWeightDefault>{});
+    // 2. Filter1x1Stride1Pad0
+    add_device_operation_instances(
+        instances,
+        device_grouped_conv_bwd_weight_dl_bf16_instances<3,
+                                                         GNDHWC,
+                                                         GKZYXC,
+                                                         GNDHWK,
+                                                         ConvBwdWeightFilter1x1Stride1Pad0>{});
+}
+
+} // namespace instance
+} // namespace device
+} // namespace tensor_operation
+} // namespace ck

library/src/tensor_operation_instance/gpu/grouped_conv3d_bwd_weight/device_grouped_conv3d_bwd_weight_dl_gndhwc_gkzyxc_gndhwk_f16_instance.cpp

+// SPDX-License-Identifier: MIT
+// Copyright (c) 2018-2023, Advanced Micro Devices, Inc. All rights reserved.
+
+#include "ck/library/tensor_operation_instance/add_device_operation_instance.hpp"
+#include "ck/library/tensor_operation_instance/gpu/grouped_conv_bwd_weight/device_grouped_conv_bwd_weight_dl_instance.hpp"
+
+namespace ck {
+namespace tensor_operation {
+namespace device {
+namespace instance {
+
+void add_device_grouped_conv3d_bwd_weight_dl_gndhwc_gkzyxc_gndhwk_f16_instances(
+    std::vector<std::unique_ptr<DeviceGroupedConvBwdWeight<3,
+                                                           GNDHWC,
+                                                           GKZYXC,
+                                                           GNDHWK,
+                                                           F16,
+                                                           F16,
+                                                           F16,
+                                                           PassThrough,
+                                                           PassThrough,
+                                                           PassThrough>>>& instances)
+{
+    // 1. Default
+    add_device_operation_instances(
+        instances,
+        device_grouped_conv_bwd_weight_dl_f16_instances<3,
+                                                        GNDHWC,
+                                                        GKZYXC,
+                                                        GNDHWK,
+                                                        ConvBwdWeightDefault>{});
+    // 2. Filter1x1Stride1Pad0
+    add_device_operation_instances(
+        instances,
+        device_grouped_conv_bwd_weight_dl_f16_instances<3,
+                                                        GNDHWC,
+                                                        GKZYXC,
+                                                        GNDHWK,
+                                                        ConvBwdWeightFilter1x1Stride1Pad0>{});
+}
+
+} // namespace instance
+} // namespace device
+} // namespace tensor_operation
+} // namespace ck

library/src/tensor_operation_instance/gpu/grouped_conv3d_bwd_weight/device_grouped_conv3d_bwd_weight_dl_gndhwc_gkzyxc_gndhwk_f32_instance.cpp

+// SPDX-License-Identifier: MIT
+// Copyright (c) 2018-2023, Advanced Micro Devices, Inc. All rights reserved.
+
+#include "ck/library/tensor_operation_instance/add_device_operation_instance.hpp"
+#include "ck/library/tensor_operation_instance/gpu/grouped_conv_bwd_weight/device_grouped_conv_bwd_weight_dl_instance.hpp"
+
+namespace ck {
+namespace tensor_operation {
+namespace device {
+namespace instance {
+
+void add_device_grouped_conv3d_bwd_weight_dl_gndhwc_gkzyxc_gndhwk_f32_instances(
+    std::vector<std::unique_ptr<DeviceGroupedConvBwdWeight<3,
+                                                           GNDHWC,
+                                                           GKZYXC,
+                                                           GNDHWK,
+                                                           F32,
+                                                           F32,
+                                                           F32,
+                                                           PassThrough,
+                                                           PassThrough,
+                                                           PassThrough>>>& instances)
+{
+    // 1. Default
+    add_device_operation_instances(
+        instances,
+        device_grouped_conv_bwd_weight_dl_f32_instances<3,
+                                                        GNDHWC,
+                                                        GKZYXC,
+                                                        GNDHWK,
+                                                        ConvBwdWeightDefault>{});
+    // 2. Filter1x1Stride1Pad0
+    add_device_operation_instances(
+        instances,
+        device_grouped_conv_bwd_weight_dl_f32_instances<3,
+                                                        GNDHWC,
+                                                        GKZYXC,
+                                                        GNDHWK,
+                                                        ConvBwdWeightFilter1x1Stride1Pad0>{});
+}
+
+} // namespace instance
+} // namespace device
+} // namespace tensor_operation
+} // namespace ck

library/src/tensor_operation_instance/gpu/grouped_conv3d_bwd_weight/device_grouped_conv3d_bwd_weight_dl_ndhwgc_gkzyxc_ndhwgk_bf16_instance.cpp

+// SPDX-License-Identifier: MIT
+// Copyright (c) 2018-2023, Advanced Micro Devices, Inc. All rights reserved.
+
+#include "ck/library/tensor_operation_instance/add_device_operation_instance.hpp"
+#include "ck/library/tensor_operation_instance/gpu/grouped_conv_bwd_weight/device_grouped_conv_bwd_weight_dl_instance.hpp"
+
+namespace ck {
+namespace tensor_operation {
+namespace device {
+namespace instance {
+
+// Compilation parameters for in[n, hi, wi, g, c] * wei[g, k, y, x, c] = out[n, ho, wo, g, k]
+void add_device_grouped_conv3d_bwd_weight_dl_ndhwgc_gkzyxc_ndhwgk_bf16_f32_bf16_instances(
+    std::vector<std::unique_ptr<DeviceGroupedConvBwdWeight<3,
+                                                           NDHWGC,
+                                                           GKZYXC,
+                                                           NDHWGK,
+                                                           BF16,
+                                                           F32,
+                                                           BF16,
+                                                           PassThrough,
+                                                           PassThrough,
+                                                           PassThrough>>>& instances)
+{
+    // 1. Default
+    add_device_operation_instances(
+        instances,
+        device_grouped_conv_bwd_weight_dl_bf16_instances<3,
+                                                         NDHWGC,
+                                                         GKZYXC,
+                                                         NDHWGK,
+                                                         ConvBwdWeightDefault>{});
+    // 2. Filter1x1Stride1Pad0
+    add_device_operation_instances(
+        instances,
+        device_grouped_conv_bwd_weight_dl_bf16_instances<3,
+                                                         NDHWGC,
+                                                         GKZYXC,
+                                                         NDHWGK,
+                                                         ConvBwdWeightFilter1x1Stride1Pad0>{});
+}
+
+} // namespace instance
+} // namespace device
+} // namespace tensor_operation
+} // namespace ck

library/src/tensor_operation_instance/gpu/grouped_conv3d_bwd_weight/device_grouped_conv3d_bwd_weight_dl_ndhwgc_gkzyxc_ndhwgk_f16_instance.cpp

+// SPDX-License-Identifier: MIT
+// Copyright (c) 2018-2023, Advanced Micro Devices, Inc. All rights reserved.
+
+#include "ck/library/tensor_operation_instance/add_device_operation_instance.hpp"
+#include "ck/library/tensor_operation_instance/gpu/grouped_conv_bwd_weight/device_grouped_conv_bwd_weight_dl_instance.hpp"
+
+namespace ck {
+namespace tensor_operation {
+namespace device {
+namespace instance {
+
+// Compilation parameters for in[n, hi, wi, g, c] * wei[g, k, y, x, c] = out[n, ho, wo, g, k]
+void add_device_grouped_conv3d_bwd_weight_dl_ndhwgc_gkzyxc_ndhwgk_f16_instances(
+    std::vector<std::unique_ptr<DeviceGroupedConvBwdWeight<3,
+                                                           NDHWGC,
+                                                           GKZYXC,
+                                                           NDHWGK,
+                                                           F16,
+                                                           F16,
+                                                           F16,
+                                                           PassThrough,
+                                                           PassThrough,
+                                                           PassThrough>>>& instances)
+{
+    // 1. Default
+    add_device_operation_instances(
+        instances,
+        device_grouped_conv_bwd_weight_dl_f16_instances<3,
+                                                        NDHWGC,
+                                                        GKZYXC,
+                                                        NDHWGK,
+                                                        ConvBwdWeightDefault>{});
+    // 2. Filter1x1Stride1Pad0
+    add_device_operation_instances(
+        instances,
+        device_grouped_conv_bwd_weight_dl_f16_instances<3,
+                                                        NDHWGC,
+                                                        GKZYXC,
+                                                        NDHWGK,
+                                                        ConvBwdWeightFilter1x1Stride1Pad0>{});
+}
+
+} // namespace instance
+} // namespace device
+} // namespace tensor_operation
+} // namespace ck

library/src/tensor_operation_instance/gpu/grouped_conv3d_bwd_weight/device_grouped_conv3d_bwd_weight_dl_ndhwgc_gkzyxc_ndhwgk_f32_instance.cpp

+// SPDX-License-Identifier: MIT
+// Copyright (c) 2018-2023, Advanced Micro Devices, Inc. All rights reserved.
+
+#include "ck/library/tensor_operation_instance/add_device_operation_instance.hpp"
+#include "ck/library/tensor_operation_instance/gpu/grouped_conv_bwd_weight/device_grouped_conv_bwd_weight_dl_instance.hpp"
+
+namespace ck {
+namespace tensor_operation {
+namespace device {
+namespace instance {
+
+// Compilation parameters for in[n, hi, wi, g, c] * wei[g, k, y, x, c] = out[n, ho, wo, g, k]
+void add_device_grouped_conv3d_bwd_weight_dl_ndhwgc_gkzyxc_ndhwgk_f32_instances(
+    std::vector<std::unique_ptr<DeviceGroupedConvBwdWeight<3,
+                                                           NDHWGC,
+                                                           GKZYXC,
+                                                           NDHWGK,
+                                                           F32,
+                                                           F32,
+                                                           F32,
+                                                           PassThrough,
+                                                           PassThrough,
+                                                           PassThrough>>>& instances)
+{
+    // 1. Default
+    add_device_operation_instances(
+        instances,
+        device_grouped_conv_bwd_weight_dl_f32_instances<3,
+                                                        NDHWGC,
+                                                        GKZYXC,
+                                                        NDHWGK,
+                                                        ConvBwdWeightDefault>{});
+    // 2. Filter1x1Stride1Pad0
+    add_device_operation_instances(
+        instances,
+        device_grouped_conv_bwd_weight_dl_f32_instances<3,
+                                                        NDHWGC,
+                                                        GKZYXC,
+                                                        NDHWGK,
+                                                        ConvBwdWeightFilter1x1Stride1Pad0>{});
+}
+
+} // namespace instance
+} // namespace device
+} // namespace tensor_operation
+} // namespace ck

library/src/tensor_operation_instance/gpu/grouped_conv3d_bwd_weight/device_grouped_conv3d_bwd_weight_xdl_gndhwc_gkzyxc_gndhwk_bf16_instance.cpp

 // SPDX-License-Identifier: MIT
 // Copyright (c) 2018-2023, Advanced Micro Devices, Inc. All rights reserved.
 
-#include <cstdlib>
-
-#include "ck/ck.hpp"
-#include "ck/tensor_operation/gpu/device/tensor_layout.hpp"
-#include "ck/tensor_operation/gpu/device/impl/device_grouped_conv_bwd_weight_xdl_cshuffle.hpp"
-#include "ck/tensor_operation/gpu/element/element_wise_operation.hpp"
-
 #include "ck/library/tensor_operation_instance/add_device_operation_instance.hpp"
 #include "ck/library/tensor_operation_instance/gpu/grouped_conv_bwd_weight/device_grouped_conv_bwd_weight_xdl_instance.hpp"
 

library/src/tensor_operation_instance/gpu/grouped_conv3d_bwd_weight/device_grouped_conv3d_bwd_weight_xdl_gndhwc_gkzyxc_gndhwk_f16_instance.cpp

 // SPDX-License-Identifier: MIT
 // Copyright (c) 2018-2023, Advanced Micro Devices, Inc. All rights reserved.
 
-#include <cstdlib>
-
-#include "ck/ck.hpp"
-#include "ck/tensor_operation/gpu/device/tensor_layout.hpp"
-#include "ck/tensor_operation/gpu/device/impl/device_grouped_conv_bwd_weight_xdl_cshuffle.hpp"
-#include "ck/tensor_operation/gpu/element/element_wise_operation.hpp"
-
 #include "ck/library/tensor_operation_instance/add_device_operation_instance.hpp"
 #include "ck/library/tensor_operation_instance/gpu/grouped_conv_bwd_weight/device_grouped_conv_bwd_weight_xdl_instance.hpp"
 

library/src/tensor_operation_instance/gpu/grouped_conv3d_bwd_weight/device_grouped_conv3d_bwd_weight_xdl_gndhwc_gkzyxc_gndhwk_f32_instance.cpp

 // SPDX-License-Identifier: MIT
 // Copyright (c) 2018-2023, Advanced Micro Devices, Inc. All rights reserved.
 
-#include <cstdlib>
-
-#include "ck/ck.hpp"
-#include "ck/tensor_operation/gpu/device/tensor_layout.hpp"
-#include "ck/tensor_operation/gpu/device/impl/device_grouped_conv_bwd_weight_xdl_cshuffle.hpp"
-#include "ck/tensor_operation/gpu/element/element_wise_operation.hpp"
-
 #include "ck/library/tensor_operation_instance/add_device_operation_instance.hpp"
 #include "ck/library/tensor_operation_instance/gpu/grouped_conv_bwd_weight/device_grouped_conv_bwd_weight_xdl_instance.hpp"