adapt conv api

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

@@ -100,15 +100,16 @@ template <index_t NDimSpatial,
           typename ELayout,
           typename ADataType,
           typename BDataType,
-          typename DsDataType,
-          typename EDataType,
           typename AccDataType,
           typename CShuffleDataType,
+          typename DsDataType,
+          typename EDataType,
           typename AElementwiseOperation,
           typename BElementwiseOperation,
           typename CDEElementwiseOperation,
           ConvolutionForwardSpecialization ConvForwardSpecialization,
           GemmSpecialization GemmSpec,
+          index_t NumGemmKPrefetchStage,
           ck::index_t BlockSize,
           ck::index_t MPerBlock,
           ck::index_t NPerBlock,
@@ -136,7 +137,6 @@ template <index_t NDimSpatial,
           index_t CShuffleNRepeatPerShuffle,
           typename CDEShuffleBlockTransferClusterLengths_MBlock_MPerBlock_NBlock_NPerBlock,
           index_t CDEShuffleBlockTransferScalarPerVector_NPerBlock,
-          index_t NumGemmKPrefetchStage   = 1,
           LoopScheduler LoopSched         = make_default_loop_scheduler(),
           ck::PipelineVersion PipelineVer = ck::PipelineVersion::v1>
 struct DeviceGroupedConvFwdMultipleD_Wmma_CShuffle