remove gemmk0 pad for output

include/ck/tensor_operation/gpu/device/device_conv2d_backward_weight_xdl_c_shuffle_nhwc_kyxc_nhwk.hpp

@@ -14,6 +14,7 @@
 #include "gridwise_gemm_xdlops_v2r4r2.hpp"
 
 #define SPLITN0_N1 1
+#define GEMMK0PAD_FOR_OUT 0
 
 namespace ck {
 namespace tensor_operation {
@@ -162,27 +163,27 @@ struct DeviceConv2dBwdWeightXdl_C_Shuffle_Input_N_Hi_Wi_C_Weight_K_Y_X_C_Output_
         const index_t GemmK0Pad = GemmKBatch * GemmK0S;
 
         const auto out_n_ho_wo_k_grid_desc =
-            make_naive_tensor_descriptor_packed(make_tuple(N, Ho, Wo, K));
+            make_naive_tensor_descriptor_packed(make_tuple(N, Ho * Wo, K));
 
         const auto out_n0_ho_wo_k_n1_grid_desc = transform_tensor_descriptor(
             out_n_ho_wo_k_grid_desc,
             make_tuple(make_unmerge_transform(make_tuple(N0, N1Number)),
-                       make_pass_through_transform(Ho),
-                       make_pass_through_transform(Wo),
+                       make_pass_through_transform(Ho * Wo),
                        make_pass_through_transform(K)),
-            make_tuple(Sequence<0>{}, Sequence<1>{}, Sequence<2>{}, Sequence<3>{}),
-            make_tuple(Sequence<0, 4>{}, Sequence<1>{}, Sequence<2>{}, Sequence<3>{})
+            make_tuple(Sequence<0>{}, Sequence<1>{}, Sequence<2>{}),
+            make_tuple(Sequence<0, 3>{}, Sequence<1>{}, Sequence<2>{})
         );
 
         const auto out_gemmk0total_gemmm_gemmk1_grid_desc = transform_tensor_descriptor(
             out_n0_ho_wo_k_n1_grid_desc,
-            make_tuple(make_merge_transform(make_tuple(N0, Ho, Wo)),
+            make_tuple(make_merge_transform(make_tuple(N0, Ho * Wo)),
                        make_pass_through_transform(K), 
                        make_pass_through_transform(N1Number)),
-            make_tuple(Sequence<0, 1, 2>{}, Sequence<3>{}, Sequence<4>{}),
+            make_tuple(Sequence<0, 1>{}, Sequence<2>{}, Sequence<3>{}),
             make_tuple(Sequence<0>{}, Sequence<1>{}, Sequence<2>{})
         );
 
+#if GEMMK0PAD_FOR_OUT
         const auto out_gemmk0pad_gemmm_gemmk1_grid_desc = transform_tensor_descriptor(
             out_gemmk0total_gemmm_gemmk1_grid_desc,
             make_tuple(make_right_pad_transform(GemmK0Total, GemmK0Pad - GemmK0Total),
@@ -198,6 +199,15 @@ struct DeviceConv2dBwdWeightXdl_C_Shuffle_Input_N_Hi_Wi_C_Weight_K_Y_X_C_Output_
                        make_pass_through_transform(N1Number)),
             make_tuple(Sequence<0>{}, Sequence<1>{}, Sequence<2>{}),
             make_tuple(Sequence<0, 1>{}, Sequence<2>{}, Sequence<3>{}));
+#else
+        const auto out_gemmkbatch_gemmk0_gemmm_gemmk1_grid_desc = transform_tensor_descriptor(
+            out_gemmk0total_gemmm_gemmk1_grid_desc,
+            make_tuple(make_unmerge_transform(make_tuple(GemmKBatch, GemmK0)),
+                       make_pass_through_transform(GemmM),
+                       make_pass_through_transform(N1Number)),
+            make_tuple(Sequence<0>{}, Sequence<1>{}, Sequence<2>{}),
+            make_tuple(Sequence<0, 1>{}, Sequence<2>{}, Sequence<3>{}));
+#endif
 #endif
         // B: input tensor
         const auto in_n_hip_wip_c_grid_desc = transform_tensor_descriptor(