Refine naming

example/27_layernorm/layernorm_fp16.cpp

@@ -27,7 +27,7 @@ using DeviceInstance =
                                                           32,  // ClusterK
                                                           1,   // SliceM
                                                           8,   // SliceK
-                                                          1,   // SrcVecDim (0=M, 1=K)
+                                                          1,   // XYVectorDim (0=M, 1=K)
                                                           8,   // SrcScalarPerVector
                                                           1,   // GammaVecDim (0=M, 1=K)
                                                           8,   // GammaScalarPerVector

example/27_layernorm/layernorm_splitk_fp16.cpp

@@ -27,7 +27,7 @@ using DeviceInstance =
                                                                 32,  // ClusterK
                                                                 1,   // SliceM
                                                                 8,   // SliceK
-                                                                1,   // SrcVecDim (0=M, 1=K)
+                                                                1,   // XYVectorDim (0=M, 1=K)
                                                                 8,   // XScalarPerVector
                                                                 1,   // GammaVecDim (0=M, 1=K)
                                                                 8,   // GammaScalarPerVector

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

@@ -221,17 +221,17 @@ struct DeviceNormalizationSplitKImpl : public DeviceNormalization<XDataType,
     template <typename DoPads, index_t MPerTile, index_t KPerTile>
     static auto MakeMeanVarDescriptor_M_K(index_t M, index_t K)
     {
-        const auto grid_desc_m_n =
+        const auto grid_desc_m_k =
             make_naive_tensor_descriptor(make_tuple(M, K), make_tuple(K, I1));
-        return PadTensorDescriptor(grid_desc_m_n, make_tuple(MPerTile, KPerTile), DoPads{});
+        return PadTensorDescriptor(grid_desc_m_k, make_tuple(MPerTile, KPerTile), DoPads{});
     }
 
     template <typename DoPads, index_t MPerTile, index_t KPerTile>
     static auto MakeCountDescriptor_M_K(index_t M, index_t K)
     {
-        const auto grid_desc_m_n =
+        const auto grid_desc_m_k =
             make_naive_tensor_descriptor(make_tuple(M, K), make_tuple(I0, I1));
-        return PadTensorDescriptor(grid_desc_m_n, make_tuple(MPerTile, KPerTile), DoPads{});
+        return PadTensorDescriptor(grid_desc_m_k, make_tuple(MPerTile, KPerTile), DoPads{});
     }
 
     using SrcGridDesc_M_K = decltype(MakeSrc2dDescriptor({1}, {1}, 1, 1));