diff --git a/library/include/ck_tile/library/tensor_operation_instance/gpu/tile_grouped_conv_bwd_weight_bf16_instances.hpp b/library/include/ck_tile/library/tensor_operation_instance/gpu/tile_grouped_conv_bwd_weight_bf16_instances.hpp
index fe197abfdb..cf67a64b8b 100644
--- a/library/include/ck_tile/library/tensor_operation_instance/gpu/tile_grouped_conv_bwd_weight_bf16_instances.hpp
+++ b/library/include/ck_tile/library/tensor_operation_instance/gpu/tile_grouped_conv_bwd_weight_bf16_instances.hpp
@@ -19,36 +19,36 @@ using tile_grouped_conv_bwd_weight_bf16_instances = std::tuple<
     //#####################################|       Num|  InLayout| WeiLayout| OutLayout| InData| WeiData| OutData|           In|         Wei|         Out|  K-block|  M-tile| N-tile | K-tile | M-warp| N-warp| K-warp| M-warp| N-warp| K-warp| Vector| Vector| Vector|  Double|                         GEMM|
     //#####################################|       Dim|          |          |          |   Type|    Type|    Type|  Elementwise| Elementwise| Elementwise|      per|        |        |        |       |       |       |   tile|   tile|   tile|   size|   size|   size|    smem|                     pipeline|
     //#####################################|   Spatial|          |          |          |       |        |        |    Operation|   Operation|   Operation|       CU|        |        |        |       |       |       |   size|   size|   size|      A|      B|      C|  buffer|                      version|
-    // GroupedConvolutionBackwardWeightInvoker<NDimSpatial,   ALayout,   BLayout,   ELayout,   BF16,    BF16,    BF16,  PassThrough, PassThrough, PassThrough,        2,      64,      64,      64,      2,      2,      1,     32,     32,     16,      2,     4,       2,   false,  CK_TILE_PIPELINE_COMPUTE_V3>,
-    // GroupedConvolutionBackwardWeightInvoker<NDimSpatial,   ALayout,   BLayout,   ELayout,   BF16,    BF16,    BF16,  PassThrough, PassThrough, PassThrough,        2,      64,      64,      64,      2,      2,      1,     32,     32,      8,      2,     4,       2,   false,  CK_TILE_PIPELINE_COMPUTE_V3>,
-    // GroupedConvolutionBackwardWeightInvoker<NDimSpatial,   ALayout,   BLayout,   ELayout,   BF16,    BF16,    BF16,  PassThrough, PassThrough, PassThrough,        2,      64,      64,      64,      2,      2,      1,     16,     16,     32,      2,     4,       2,   false,  CK_TILE_PIPELINE_COMPUTE_V3>,
-    // GroupedConvolutionBackwardWeightInvoker<NDimSpatial,   ALayout,   BLayout,   ELayout,   BF16,    BF16,    BF16,  PassThrough, PassThrough, PassThrough,        2,      64,      64,      64,      2,      2,      1,     16,     16,     32,      4,     4,       4,   false,  CK_TILE_PIPELINE_COMPUTE_V3>,
-    // GroupedConvolutionBackwardWeightInvoker<NDimSpatial,   ALayout,   BLayout,   ELayout,   BF16,    BF16,    BF16,  PassThrough, PassThrough, PassThrough,        2,      64,      64,      64,      2,      2,      1,     16,     16,     32,      8,     8,       8,   false,  CK_TILE_PIPELINE_COMPUTE_V3>,
-    // GroupedConvolutionBackwardWeightInvoker<NDimSpatial,   ALayout,   BLayout,   ELayout,   BF16,    BF16,    BF16,  PassThrough, PassThrough, PassThrough,        1,      64,      64,      64,      1,      1,      1,     32,     32,      8,      2,     2,       2,   false,  CK_TILE_PIPELINE_COMPUTE_V3>,
-    // GroupedConvolutionBackwardWeightInvoker<NDimSpatial,   ALayout,   BLayout,   ELayout,   BF16,    BF16,    BF16,  PassThrough, PassThrough, PassThrough,        1,      64,      64,      64,      1,      1,      1,     32,     32,     16,      2,     2,       2,   false,  CK_TILE_PIPELINE_COMPUTE_V3>,
-    // GroupedConvolutionBackwardWeightInvoker<NDimSpatial,   ALayout,   BLayout,   ELayout,   BF16,    BF16,    BF16,  PassThrough, PassThrough, PassThrough,        1,      64,      64,      64,      1,      1,      1,     16,     16,     32,      2,     2,       2,   false,  CK_TILE_PIPELINE_COMPUTE_V3>,
+    GroupedConvolutionBackwardWeightInvoker<NDimSpatial,   ALayout,   BLayout,   ELayout,   BF16,    BF16,    BF16,  PassThrough, PassThrough, PassThrough,        2,      64,      64,      64,      2,      2,      1,     32,     32,     16,      2,     4,       2,   false,  CK_TILE_PIPELINE_COMPUTE_V3>,
+    GroupedConvolutionBackwardWeightInvoker<NDimSpatial,   ALayout,   BLayout,   ELayout,   BF16,    BF16,    BF16,  PassThrough, PassThrough, PassThrough,        2,      64,      64,      64,      2,      2,      1,     32,     32,      8,      2,     4,       2,   false,  CK_TILE_PIPELINE_COMPUTE_V3>,
+    GroupedConvolutionBackwardWeightInvoker<NDimSpatial,   ALayout,   BLayout,   ELayout,   BF16,    BF16,    BF16,  PassThrough, PassThrough, PassThrough,        2,      64,      64,      64,      2,      2,      1,     16,     16,     32,      2,     4,       2,   false,  CK_TILE_PIPELINE_COMPUTE_V3>,
+    GroupedConvolutionBackwardWeightInvoker<NDimSpatial,   ALayout,   BLayout,   ELayout,   BF16,    BF16,    BF16,  PassThrough, PassThrough, PassThrough,        2,      64,      64,      64,      2,      2,      1,     16,     16,     32,      4,     4,       4,   false,  CK_TILE_PIPELINE_COMPUTE_V3>,
+    GroupedConvolutionBackwardWeightInvoker<NDimSpatial,   ALayout,   BLayout,   ELayout,   BF16,    BF16,    BF16,  PassThrough, PassThrough, PassThrough,        2,      64,      64,      64,      2,      2,      1,     16,     16,     32,      8,     8,       8,   false,  CK_TILE_PIPELINE_COMPUTE_V3>,
+    GroupedConvolutionBackwardWeightInvoker<NDimSpatial,   ALayout,   BLayout,   ELayout,   BF16,    BF16,    BF16,  PassThrough, PassThrough, PassThrough,        1,      64,      64,      64,      1,      1,      1,     32,     32,      8,      2,     2,       2,   false,  CK_TILE_PIPELINE_COMPUTE_V3>,
+    GroupedConvolutionBackwardWeightInvoker<NDimSpatial,   ALayout,   BLayout,   ELayout,   BF16,    BF16,    BF16,  PassThrough, PassThrough, PassThrough,        1,      64,      64,      64,      1,      1,      1,     32,     32,     16,      2,     2,       2,   false,  CK_TILE_PIPELINE_COMPUTE_V3>,
+    GroupedConvolutionBackwardWeightInvoker<NDimSpatial,   ALayout,   BLayout,   ELayout,   BF16,    BF16,    BF16,  PassThrough, PassThrough, PassThrough,        1,      64,      64,      64,      1,      1,      1,     16,     16,     32,      2,     2,       2,   false,  CK_TILE_PIPELINE_COMPUTE_V3>,
 
-    // GroupedConvolutionBackwardWeightInvoker<NDimSpatial,   ALayout,   BLayout,   ELayout,   BF16,    BF16,    BF16,  PassThrough, PassThrough, PassThrough,        1,      16,      16,      64,      1,      1,      1,     16,     16,     32,      8,     8,       2,   false,  CK_TILE_PIPELINE_COMPUTE_V3>,
-    // GroupedConvolutionBackwardWeightInvoker<NDimSpatial,   ALayout,   BLayout,   ELayout,   BF16,    BF16,    BF16,  PassThrough, PassThrough, PassThrough,        2,      16,      16,      64,      1,      1,      1,     16,     16,     32,      8,     8,       2,   false,  CK_TILE_PIPELINE_COMPUTE_V3>,
-    // GroupedConvolutionBackwardWeightInvoker<NDimSpatial,   ALayout,   BLayout,   ELayout,   BF16,    BF16,    BF16,  PassThrough, PassThrough, PassThrough,        1,      16,      16,      64,      1,      1,      1,     16,     16,     32,      4,     4,       2,   false,  CK_TILE_PIPELINE_COMPUTE_V3>,
-    // GroupedConvolutionBackwardWeightInvoker<NDimSpatial,   ALayout,   BLayout,   ELayout,   BF16,    BF16,    BF16,  PassThrough, PassThrough, PassThrough,        2,      16,      16,      64,      1,      1,      1,     16,     16,     32,      4,     4,       2,   false,  CK_TILE_PIPELINE_COMPUTE_V3>,
-    // GroupedConvolutionBackwardWeightInvoker<NDimSpatial,   ALayout,   BLayout,   ELayout,   BF16,    BF16,    BF16,  PassThrough, PassThrough, PassThrough,        1,      16,      16,      64,      1,      1,      1,     16,     16,     32,      4,     4,       4,   false,  CK_TILE_PIPELINE_COMPUTE_V3>,
-    // GroupedConvolutionBackwardWeightInvoker<NDimSpatial,   ALayout,   BLayout,   ELayout,   BF16,    BF16,    BF16,  PassThrough, PassThrough, PassThrough,        2,      16,      16,      64,      1,      1,      1,     16,     16,     32,      4,     4,       4,   false,  CK_TILE_PIPELINE_COMPUTE_V3>,
+    GroupedConvolutionBackwardWeightInvoker<NDimSpatial,   ALayout,   BLayout,   ELayout,   BF16,    BF16,    BF16,  PassThrough, PassThrough, PassThrough,        1,      16,      16,      64,      1,      1,      1,     16,     16,     32,      8,     8,       2,   false,  CK_TILE_PIPELINE_COMPUTE_V3>,
+    GroupedConvolutionBackwardWeightInvoker<NDimSpatial,   ALayout,   BLayout,   ELayout,   BF16,    BF16,    BF16,  PassThrough, PassThrough, PassThrough,        2,      16,      16,      64,      1,      1,      1,     16,     16,     32,      8,     8,       2,   false,  CK_TILE_PIPELINE_COMPUTE_V3>,
+    GroupedConvolutionBackwardWeightInvoker<NDimSpatial,   ALayout,   BLayout,   ELayout,   BF16,    BF16,    BF16,  PassThrough, PassThrough, PassThrough,        1,      16,      16,      64,      1,      1,      1,     16,     16,     32,      4,     4,       2,   false,  CK_TILE_PIPELINE_COMPUTE_V3>,
+    GroupedConvolutionBackwardWeightInvoker<NDimSpatial,   ALayout,   BLayout,   ELayout,   BF16,    BF16,    BF16,  PassThrough, PassThrough, PassThrough,        2,      16,      16,      64,      1,      1,      1,     16,     16,     32,      4,     4,       2,   false,  CK_TILE_PIPELINE_COMPUTE_V3>,
+    GroupedConvolutionBackwardWeightInvoker<NDimSpatial,   ALayout,   BLayout,   ELayout,   BF16,    BF16,    BF16,  PassThrough, PassThrough, PassThrough,        1,      16,      16,      64,      1,      1,      1,     16,     16,     32,      4,     4,       4,   false,  CK_TILE_PIPELINE_COMPUTE_V3>,
+    GroupedConvolutionBackwardWeightInvoker<NDimSpatial,   ALayout,   BLayout,   ELayout,   BF16,    BF16,    BF16,  PassThrough, PassThrough, PassThrough,        2,      16,      16,      64,      1,      1,      1,     16,     16,     32,      4,     4,       4,   false,  CK_TILE_PIPELINE_COMPUTE_V3>,
 
-    // GroupedConvolutionBackwardWeightInvoker<NDimSpatial,   ALayout,   BLayout,   ELayout,   BF16,    BF16,    BF16,  PassThrough, PassThrough, PassThrough,        2,      64,      64,      64,      2,      2,      1,     32,     32,     16,      2,     4,       2,    true,  CK_TILE_PIPELINE_COMPUTE_V4>,
-    // GroupedConvolutionBackwardWeightInvoker<NDimSpatial,   ALayout,   BLayout,   ELayout,   BF16,    BF16,    BF16,  PassThrough, PassThrough, PassThrough,        2,      64,      64,      64,      2,      2,      1,     32,     32,      8,      2,     4,       2,    true,  CK_TILE_PIPELINE_COMPUTE_V4>,
-    // GroupedConvolutionBackwardWeightInvoker<NDimSpatial,   ALayout,   BLayout,   ELayout,   BF16,    BF16,    BF16,  PassThrough, PassThrough, PassThrough,        2,      64,      64,      64,      2,      2,      1,     16,     16,     32,      2,     4,       2,    true,  CK_TILE_PIPELINE_COMPUTE_V4>,
-    // GroupedConvolutionBackwardWeightInvoker<NDimSpatial,   ALayout,   BLayout,   ELayout,   BF16,    BF16,    BF16,  PassThrough, PassThrough, PassThrough,        2,      64,      64,      64,      2,      2,      1,     16,     16,     32,      4,     4,       4,    true,  CK_TILE_PIPELINE_COMPUTE_V4>,
-    // GroupedConvolutionBackwardWeightInvoker<NDimSpatial,   ALayout,   BLayout,   ELayout,   BF16,    BF16,    BF16,  PassThrough, PassThrough, PassThrough,        2,      64,      64,      64,      2,      2,      1,     16,     16,     32,      8,     8,       8,    true,  CK_TILE_PIPELINE_COMPUTE_V4>,
-    // GroupedConvolutionBackwardWeightInvoker<NDimSpatial,   ALayout,   BLayout,   ELayout,   BF16,    BF16,    BF16,  PassThrough, PassThrough, PassThrough,        1,      64,      64,      64,      1,      1,      1,     32,     32,      8,      2,     2,       2,    true,  CK_TILE_PIPELINE_COMPUTE_V4>,
-    // GroupedConvolutionBackwardWeightInvoker<NDimSpatial,   ALayout,   BLayout,   ELayout,   BF16,    BF16,    BF16,  PassThrough, PassThrough, PassThrough,        1,      64,      64,      64,      1,      1,      1,     32,     32,     16,      2,     2,       2,    true,  CK_TILE_PIPELINE_COMPUTE_V4>,
-    // GroupedConvolutionBackwardWeightInvoker<NDimSpatial,   ALayout,   BLayout,   ELayout,   BF16,    BF16,    BF16,  PassThrough, PassThrough, PassThrough,        1,      64,      64,      64,      1,      1,      1,     16,     16,     32,      2,     2,       2,    true,  CK_TILE_PIPELINE_COMPUTE_V4>,
+    GroupedConvolutionBackwardWeightInvoker<NDimSpatial,   ALayout,   BLayout,   ELayout,   BF16,    BF16,    BF16,  PassThrough, PassThrough, PassThrough,        2,      64,      64,      64,      2,      2,      1,     32,     32,     16,      2,     4,       2,    true,  CK_TILE_PIPELINE_COMPUTE_V4>,
+    GroupedConvolutionBackwardWeightInvoker<NDimSpatial,   ALayout,   BLayout,   ELayout,   BF16,    BF16,    BF16,  PassThrough, PassThrough, PassThrough,        2,      64,      64,      64,      2,      2,      1,     32,     32,      8,      2,     4,       2,    true,  CK_TILE_PIPELINE_COMPUTE_V4>,
+    GroupedConvolutionBackwardWeightInvoker<NDimSpatial,   ALayout,   BLayout,   ELayout,   BF16,    BF16,    BF16,  PassThrough, PassThrough, PassThrough,        2,      64,      64,      64,      2,      2,      1,     16,     16,     32,      2,     4,       2,    true,  CK_TILE_PIPELINE_COMPUTE_V4>,
+    GroupedConvolutionBackwardWeightInvoker<NDimSpatial,   ALayout,   BLayout,   ELayout,   BF16,    BF16,    BF16,  PassThrough, PassThrough, PassThrough,        2,      64,      64,      64,      2,      2,      1,     16,     16,     32,      4,     4,       4,    true,  CK_TILE_PIPELINE_COMPUTE_V4>,
+    GroupedConvolutionBackwardWeightInvoker<NDimSpatial,   ALayout,   BLayout,   ELayout,   BF16,    BF16,    BF16,  PassThrough, PassThrough, PassThrough,        2,      64,      64,      64,      2,      2,      1,     16,     16,     32,      8,     8,       8,    true,  CK_TILE_PIPELINE_COMPUTE_V4>,
+    GroupedConvolutionBackwardWeightInvoker<NDimSpatial,   ALayout,   BLayout,   ELayout,   BF16,    BF16,    BF16,  PassThrough, PassThrough, PassThrough,        1,      64,      64,      64,      1,      1,      1,     32,     32,      8,      2,     2,       2,    true,  CK_TILE_PIPELINE_COMPUTE_V4>,
+    GroupedConvolutionBackwardWeightInvoker<NDimSpatial,   ALayout,   BLayout,   ELayout,   BF16,    BF16,    BF16,  PassThrough, PassThrough, PassThrough,        1,      64,      64,      64,      1,      1,      1,     32,     32,     16,      2,     2,       2,    true,  CK_TILE_PIPELINE_COMPUTE_V4>,
+    GroupedConvolutionBackwardWeightInvoker<NDimSpatial,   ALayout,   BLayout,   ELayout,   BF16,    BF16,    BF16,  PassThrough, PassThrough, PassThrough,        1,      64,      64,      64,      1,      1,      1,     16,     16,     32,      2,     2,       2,    true,  CK_TILE_PIPELINE_COMPUTE_V4>,
 
-    // GroupedConvolutionBackwardWeightInvoker<NDimSpatial,   ALayout,   BLayout,   ELayout,   BF16,    BF16,    BF16,  PassThrough, PassThrough, PassThrough,        1,      16,      16,      64,      1,      1,      1,     16,     16,     32,      8,     8,       2,    true,  CK_TILE_PIPELINE_COMPUTE_V4>,
-    // GroupedConvolutionBackwardWeightInvoker<NDimSpatial,   ALayout,   BLayout,   ELayout,   BF16,    BF16,    BF16,  PassThrough, PassThrough, PassThrough,        2,      16,      16,      64,      1,      1,      1,     16,     16,     32,      8,     8,       2,    true,  CK_TILE_PIPELINE_COMPUTE_V4>,
-    // GroupedConvolutionBackwardWeightInvoker<NDimSpatial,   ALayout,   BLayout,   ELayout,   BF16,    BF16,    BF16,  PassThrough, PassThrough, PassThrough,        1,      16,      16,      64,      1,      1,      1,     16,     16,     32,      4,     4,       2,    true,  CK_TILE_PIPELINE_COMPUTE_V4>,
-    // GroupedConvolutionBackwardWeightInvoker<NDimSpatial,   ALayout,   BLayout,   ELayout,   BF16,    BF16,    BF16,  PassThrough, PassThrough, PassThrough,        2,      16,      16,      64,      1,      1,      1,     16,     16,     32,      4,     4,       2,    true,  CK_TILE_PIPELINE_COMPUTE_V4>,
-    // GroupedConvolutionBackwardWeightInvoker<NDimSpatial,   ALayout,   BLayout,   ELayout,   BF16,    BF16,    BF16,  PassThrough, PassThrough, PassThrough,        1,      16,      16,      64,      1,      1,      1,     16,     16,     32,      4,     4,       4,    true,  CK_TILE_PIPELINE_COMPUTE_V4>,
+    GroupedConvolutionBackwardWeightInvoker<NDimSpatial,   ALayout,   BLayout,   ELayout,   BF16,    BF16,    BF16,  PassThrough, PassThrough, PassThrough,        1,      16,      16,      64,      1,      1,      1,     16,     16,     32,      8,     8,       2,    true,  CK_TILE_PIPELINE_COMPUTE_V4>,
+    GroupedConvolutionBackwardWeightInvoker<NDimSpatial,   ALayout,   BLayout,   ELayout,   BF16,    BF16,    BF16,  PassThrough, PassThrough, PassThrough,        2,      16,      16,      64,      1,      1,      1,     16,     16,     32,      8,     8,       2,    true,  CK_TILE_PIPELINE_COMPUTE_V4>,
+    GroupedConvolutionBackwardWeightInvoker<NDimSpatial,   ALayout,   BLayout,   ELayout,   BF16,    BF16,    BF16,  PassThrough, PassThrough, PassThrough,        1,      16,      16,      64,      1,      1,      1,     16,     16,     32,      4,     4,       2,    true,  CK_TILE_PIPELINE_COMPUTE_V4>,
+    GroupedConvolutionBackwardWeightInvoker<NDimSpatial,   ALayout,   BLayout,   ELayout,   BF16,    BF16,    BF16,  PassThrough, PassThrough, PassThrough,        2,      16,      16,      64,      1,      1,      1,     16,     16,     32,      4,     4,       2,    true,  CK_TILE_PIPELINE_COMPUTE_V4>,
+    GroupedConvolutionBackwardWeightInvoker<NDimSpatial,   ALayout,   BLayout,   ELayout,   BF16,    BF16,    BF16,  PassThrough, PassThrough, PassThrough,        1,      16,      16,      64,      1,      1,      1,     16,     16,     32,      4,     4,       4,    true,  CK_TILE_PIPELINE_COMPUTE_V4>,
     GroupedConvolutionBackwardWeightInvoker<NDimSpatial,   ALayout,   BLayout,   ELayout,   BF16,    BF16,    BF16,  PassThrough, PassThrough, PassThrough,        2,      16,      16,      64,      1,      1,      1,     16,     16,     32,      2,     2,       2,    true,  CK_TILE_PIPELINE_COMPUTE_V4>
     // clang-format on
 >;