Merge branch 'feature/integrage-karg-simplification-pr' into...

Merge branch 'feature/integrage-karg-simplification-pr' into feature/simplify-karg-for-device-gemm-xdl

include/ck/tensor_operation/gpu/device/impl/device_cgemm_4gemm_xdl_cshuffle.hpp

@@ -285,7 +285,7 @@ struct DeviceCGemm_4Gemm_Xdl_CShuffle
             if(GridwiseGemm::CalculateHasMainKBlockLoop(K))
             {
                 const auto kernel =
-                    kernel_gemm_xdl_cshuffle_v2<GridwiseGemm, ADataType, CDataType, true>;
+                    kernel_gemm_xdl_cshuffle_v1<GridwiseGemm, ADataType, CDataType, true>;
 
                 ave_time += launch_and_time_kernel(stream_config,
                                                    kernel,
@@ -358,7 +358,7 @@ struct DeviceCGemm_4Gemm_Xdl_CShuffle
             else
             {
                 const auto kernel =
-                    kernel_gemm_xdl_cshuffle_v2<GridwiseGemm, ADataType, CDataType, false>;
+                    kernel_gemm_xdl_cshuffle_v1<GridwiseGemm, ADataType, CDataType, false>;
 
                 ave_time += launch_and_time_kernel(stream_config,
                                                    kernel,

include/ck/tensor_operation/gpu/grid/gridwise_gemm_xdl_cshuffle_v1.hpp

@@ -43,7 +43,7 @@ __global__ void
 #if CK_USE_LAUNCH_BOUNDS
     __launch_bounds__(CK_MAX_THREAD_PER_BLOCK, CK_MIN_BLOCK_PER_CU)
 #endif
-        kernel_gemm_xdl_cshuffle_v2(const FloatAB* __restrict__ p_a_grid,
+        kernel_gemm_xdl_cshuffle_v1(const FloatAB* __restrict__ p_a_grid,
                                     const FloatAB* __restrict__ p_b_grid,
                                     FloatC* __restrict__ p_c_grid,
                                     typename GridwiseGemm::Problem problem)