diff --git a/composable_kernel/include/tensor_operation/blockwise_dynamic_tensor_slice_transfer.hpp b/composable_kernel/include/tensor_operation/blockwise_dynamic_tensor_slice_transfer.hpp
index ca116ef17e..36de380719 100644
--- a/composable_kernel/include/tensor_operation/blockwise_dynamic_tensor_slice_transfer.hpp
+++ b/composable_kernel/include/tensor_operation/blockwise_dynamic_tensor_slice_transfer.hpp
@@ -89,6 +89,16 @@ struct BlockwiseDynamicTensorSliceTransfer_v4
         }
     }
 
+    template <typename SrcBuffer>
+    __device__ void RunRead(const SrcDesc& src_desc, const SrcBuffer& src_buf)
+    {
+        if(BlockSize == thread_cluster_desc_.GetElementSize() or
+           get_thread_local_1d_id() < thread_cluster_desc_.GetElementSize())
+        {
+            threadwise_transfer_.RunRead(src_desc, src_buf);
+        }
+    }
+
     template <typename DstBuffer>
     __device__ void RunWrite(const DstDesc& dst_desc, DstBuffer& dst_buf)
     {
diff --git a/driver/include/device_dynamic_convolution_forward_implicit_gemm_v4r4r2_xdlops_nchw_kcyx_nkhw.hpp b/driver/include/device_dynamic_convolution_forward_implicit_gemm_v4r4r2_xdlops_nchw_kcyx_nkhw.hpp
index 5b6a94a0b7..4f531b55ae 100644
--- a/driver/include/device_dynamic_convolution_forward_implicit_gemm_v4r4r2_xdlops_nchw_kcyx_nkhw.hpp
+++ b/driver/include/device_dynamic_convolution_forward_implicit_gemm_v4r4r2_xdlops_nchw_kcyx_nkhw.hpp
@@ -150,7 +150,7 @@ void device_dynamic_convolution_forward_implicit_gemm_v4r4r2_xdlops_nchw_kcyx_nk
 
     constexpr index_t GemmMPerWave = 64;
     constexpr index_t GemmNPerWave = 64;
-    constexpr index_t GemmK1    = 8;
+    constexpr index_t GemmK1       = 8;
 
     constexpr index_t MRepeat = 1;
     constexpr index_t NRepeat = 1;
@@ -164,7 +164,7 @@ void device_dynamic_convolution_forward_implicit_gemm_v4r4r2_xdlops_nchw_kcyx_nk
     using GemmBBlockTransferThreadSliceLengths_GemmK0_GemmN_GemmK1   = Sequence<1, 4, 4>;
     using GemmBBlockTransferThreadClusterLengths_GemmK0_GemmN_GemmK1 = Sequence<4, 32, 2>;
 
-    constexpr index_t GemmBBlockTransferSrcScalarPerVector_GemmN = 1;
+    constexpr index_t GemmBBlockTransferSrcScalarPerVector_GemmN  = 1;
     constexpr index_t GemmBBlockTransferDstScalarPerVector_GemmK1 = 4;
 
     constexpr index_t GemmCThreadTransferDstScalarPerVector = 1;