Reuse the existing implementation

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

@@ -125,14 +125,17 @@ struct BlockToCTileMap_M00_N0_M01Adapt
     {
     }
 
-    __host__ constexpr index_t CalculateGridSize(const CGridDesc_M_N& c_grid_desc_m_n) const
+    __host__ static constexpr index_t CalculateGridSize(index_t M, index_t N)
     {
-        const auto M0 = math::integer_divide_ceil(c_grid_desc_m_n.GetLength(I0), MPerBlock);
-        const auto N0 = math::integer_divide_ceil(c_grid_desc_m_n.GetLength(I1), NPerBlock);
+        const auto M0 = math::integer_divide_ceil(M, MPerBlock);
+        const auto N0 = math::integer_divide_ceil(N, NPerBlock);
 
-        const index_t grid_size = M0 * N0;
+        return M0 * N0;
+    }
 
-        return grid_size;
+    __host__ constexpr index_t CalculateGridSize(const CGridDesc_M_N& c_grid_desc_m_n) const
+    {
+        return CalculateGridSize(c_grid_desc_m_n.GetLength(I0), c_grid_desc_m_n.GetLength(I1));
     }
 
     template <typename TopIdx>

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

@@ -144,8 +144,7 @@ struct GridwiseGemm_k0mk1_k0nk1_mn_xdl_cshuffle_v1
     __host__ __device__ static auto CalculateGridSize(index_t M, index_t N)
     {
         // reference the implementation of class 'BlockToCTileMap_M00_N0_M01Adapt'
-        return std::make_tuple(
-            INTEGER_DIVIDE_CEIL(M, MPerBlock) * INTEGER_DIVIDE_CEIL(N, NPerBlock), 1, 1);
+        return std::make_tuple(DefaultBlock2CTileMap::CalculateGridSize(M, N), 1, 1);
     }
 
     __host__ __device__ static auto CalculateMPadded(index_t M)