fix

composable_kernel/include/tensor_operation/threadwise_dynamic_tensor_slice_transfer.hpp

@@ -201,7 +201,7 @@ struct ThreadwiseDynamicTensorSliceTransfer_v1r3
 #else
                 if(is_dst_valid)
                 {
-                    *reinterpret_cast<dst_vector_t*>(&(p_dst[dst_slice_origin_coord_.GetOffset])) =
+                    *reinterpret_cast<dst_vector_t*>(&(p_dst[dst_slice_origin_coord_.GetOffset()])) =
                         dst_vector.Vector();
                 }
 #endif
@@ -210,7 +210,7 @@ struct ThreadwiseDynamicTensorSliceTransfer_v1r3
             {
                 if(is_dst_valid)
                 {
-                    *reinterpret_cast<dst_vector_t*>(&(p_dst[dst_slice_origin_coord_.GetOffset])) =
+                    *reinterpret_cast<dst_vector_t*>(&(p_dst[dst_slice_origin_coord_.GetOffset()])) =
                         dst_vector.Vector();
                 }
             }

driver/include/device_dynamic_convolution_forward_implicit_gemm_v4r4_nchw_kcyx_nkhw.hpp

@@ -39,7 +39,7 @@ void device_dynamic_convolution_forward_implicit_gemm_v4r4_nchw_kcyx_nkhw(InDesc
     wei_kcyx_device_buf.ToDevice(wei_kcyx.mData.data());
     out_nkhw_device_buf.ToDevice(out_nkhw.mData.data());
 
-#if 0
+#if 1
     // run-time variables
     const auto in_n_c_hi_wi_desc =
         make_dynamic_naive_tensor_descriptor_packed_v2(to_multi_index(InDesc::GetLengths()));