Minor stylistic fixes

example/20_grouped_conv_bwd_weight/grouped_conv_bwd_weight_dl_fp16.cpp

@@ -91,6 +91,7 @@ int main(int argc, char* argv[])
     case 1: return !run_grouped_conv_bwd_weight<1>(config, conv_param);
     case 2: return !run_grouped_conv_bwd_weight<2>(config, conv_param);
     case 3: return !run_grouped_conv_bwd_weight<3>(config, conv_param);
+    default: break;
     }
 
     return 1;

example/20_grouped_conv_bwd_weight/grouped_conv_bwd_weight_wmma_fp16.cpp

@@ -81,6 +81,7 @@ int main(int argc, char* argv[])
     switch(conv_param.num_dim_spatial_)
     {
     case 3: return !run_grouped_conv_bwd_weight<3>(config, conv_param);
+    default: break;
     }
 
     return 1;

example/20_grouped_conv_bwd_weight/grouped_conv_bwd_weight_xdl_bf16.cpp

@@ -93,6 +93,7 @@ int main(int argc, char* argv[])
     case 1: return !run_grouped_conv_bwd_weight<1>(config, conv_param);
     case 2: return !run_grouped_conv_bwd_weight<2>(config, conv_param);
     case 3: return !run_grouped_conv_bwd_weight<3>(config, conv_param);
+    default: break;
     }
 
     return 1;

example/20_grouped_conv_bwd_weight/grouped_conv_bwd_weight_xdl_fp16.cpp

@@ -92,6 +92,7 @@ int main(int argc, char* argv[])
     case 1: return !run_grouped_conv_bwd_weight<1>(config, conv_param);
     case 2: return !run_grouped_conv_bwd_weight<2>(config, conv_param);
     case 3: return !run_grouped_conv_bwd_weight<3>(config, conv_param);
+    default: break;
     }
 
     return 1;

example/20_grouped_conv_bwd_weight/grouped_conv_bwd_weight_xdl_fp16_comp_bf8_fp8.cpp

@@ -98,6 +98,7 @@ int main(int argc, char* argv[])
     case 1: return !run_grouped_conv_bwd_weight<1>(config, conv_param);
     case 2: return !run_grouped_conv_bwd_weight<2>(config, conv_param);
     case 3: return !run_grouped_conv_bwd_weight<3>(config, conv_param);
+    default: break;
     }
 
     return 1;

example/20_grouped_conv_bwd_weight/run_grouped_conv_bwd_weight_example.inc

@@ -5,7 +5,7 @@ template <ck::index_t NDimSpatial>
 bool run_grouped_conv_bwd_weight(const ExecutionConfig& config,
                                  const ck::utils::conv::ConvParam& conv_param)
 {
-    // Dl and WMMA ops doesn't support split_k > 1
+    // Dl and WMMA ops don't support split_k > 1
     constexpr ck::index_t split_k = 1;
 
     const auto in_g_n_c_wis_desc =

include/ck/tensor_operation/gpu/device/impl/device_grouped_conv_bwd_weight_wmma_cshuffle.hpp

@@ -61,9 +61,10 @@ template <index_t NDimSpatial,
           index_t CShuffleNRepeatPerShuffle,
           typename CShuffleBlockTransferClusterLengths_MBlock_MPerBlock_NBlock_NPerBlock,
           index_t CShuffleBlockTransferScalarPerVector_NPerBlock,
-          index_t NumGemmKPrefetchStage   = 1,
-          LoopScheduler LoopSched         = make_default_loop_scheduler(),
-          ck::PipelineVersion PipelineVer = ck::PipelineVersion::v1>
+          index_t NumGemmKPrefetchStage                        = 1,
+          LoopScheduler LoopSched                              = make_default_loop_scheduler(),
+          ck::PipelineVersion PipelineVer                      = ck::PipelineVersion::v1,
+          typename ck::enable_if<NDimSpatial == 3, bool>::type = false>
 struct DeviceGroupedConvBwdWeight_Wmma_CShuffle
     : public DeviceGroupedConvBwdWeight<NDimSpatial,
                                         InLayout,
@@ -332,7 +333,7 @@ struct DeviceGroupedConvBwdWeight_Wmma_CShuffle
                 make_tuple(Sequence<0>{}, Sequence<1>{}),
                 make_tuple(Sequence<0, 2>{}, Sequence<1>{}));
 
-            // Padd
+            // Pad
             const auto out_gemmkbatch_gemmk0_gemmm_gemmk1_pad_grid_desc =
                 transform_tensor_descriptor(
                     out_gemmkbatch_gemmk0_gemmm_gemmk1_grid_desc,
@@ -362,7 +363,7 @@ struct DeviceGroupedConvBwdWeight_Wmma_CShuffle
                               in_gemmkbatch_gemmk0_gemmn_gemmk1_pad_grid_desc,
                               wei_gemmm_gemmn_pad_grid_desc);
         }
-    } // function end
+    }
 
     template <index_t NDim, typename ck::enable_if<NDim == 3, bool>::type = false>
     static auto GetABCGridDesc()
@@ -716,14 +717,7 @@ struct DeviceGroupedConvBwdWeight_Wmma_CShuffle
             return false;
         }
 
-        if constexpr(NDimSpatial == 3)
-        {
-            if constexpr(!(is_NDHWGK_GKZYXC_NDHWGC || is_GNDHWK_GKZYXC_GNDHWC))
-            {
-                return false;
-            }
-        }
-        else
+        if constexpr(!(is_NDHWGK_GKZYXC_NDHWGC || is_GNDHWK_GKZYXC_GNDHWC))
         {
             return false;
         }
@@ -731,7 +725,7 @@ struct DeviceGroupedConvBwdWeight_Wmma_CShuffle
         if constexpr(ConvBackwardWeightSpecialization ==
                      ConvolutionBackwardWeightSpecialization::Filter1x1Stride1Pad0)
         {
-            // check if it's 1x1, stride=1 pad = 0 conv
+            // check if it's a 1x1 convolution with stride=1 and no padding
             for(int i = 0; i < NDimSpatial; i++)
             {
                 if(!(arg.filter_spatial_lengths_[i] == 1 && arg.conv_filter_strides_[i] == 1 &&

test/grouped_convnd_bwd_weight/test_grouped_convnd_bwd_weight.cpp

@@ -36,7 +36,7 @@ class TestGroupedConvndBwdWeight : public ::testing::Test
     {
         // Odd K or C values are supported only by DL and WMMA
         // kernels (only applies to fp16)
-        // DL kernel currently supports only `split_k=1`
+        // DL and WMMA kernels currently support only `split_k=1`
         if constexpr(std::is_same_v<InDataType, ck::half_t>)
         {
             if(split_k != 1 && (params.K_ % 2 != 0 || params.C_ % 2 != 0))

test/grouped_convnd_bwd_weight/test_grouped_convnd_bwd_weight_interface_wmma.cpp

@@ -137,13 +137,13 @@ TYPED_TEST_SUITE(TestGroupedConvndBwdWeightDefault3d, KernelTypes3d);
 
 TYPED_TEST(TestGroupedConvndBwdWeightFilter1x13d, SpecializationCheck)
 {
-    // Check filter 3,3 instead of 1,1
+    // Check filter 3x3x3 instead of 1x1x1
     this->conv_param = {
         3, 2, 4, 192, 192, {3, 3, 3}, {28, 28, 28}, {1, 1, 1}, {1, 1, 1}, {0, 0, 0}, {0, 0, 0}};
     bool is_supported = this->template Run<1>();
     EXPECT_FALSE(is_supported);
 
-    // Check strides 2,2 instead of 1,1
+    // Check strides 2x2x2 instead of 1x1x1
     this->conv_param = {
         3, 2, 4, 192, 192, {1, 1, 1}, {28, 28, 28}, {2, 2, 2}, {1, 1, 1}, {0, 0, 0}, {0, 0, 0}};
     is_supported = this->template Run<1>();