Gemm+Reduce Fusion (#128)

* add gridwise gemm v4r1

* rename

* adding gemm+reduce

* adding gemm+reduce

* adding gemm+reduce

* adding gemm+reduce

* use sfc in shuffling

* remove hardcode

* remove hardcode

* refactor

* fix build

* adding gemm+reduce

* adding gemm+reduce

* adding gemm+reduce

* adding gemm+reduce

* adding gemm+reduce

* format

* clean

* adding gemm+reduce

* adding profiler for gemm+reduce

* adding gemm+reduce profiler

* fix build

* clean up

* gemm+reduce

* fix build

* update DeviceGemm_Xdl_CShuffle; update enum to enum class

* clean up

* add test for gemm+reduce

* clean up

* refactor

* fix build

* fix build

include/ck/tensor_operation/gpu/element/element_wise_operation.hpp

@@ -1,6 +1,4 @@
-#ifndef CK_ELEMENT_WISE_OPERATION_HPP
-#define CK_ELEMENT_WISE_OPERATION_HPP
-
+#pragma once
 #include "data_type.hpp"
 
 namespace ck {
@@ -365,4 +363,3 @@ struct UnarySqrt<double, double>
 } // namespace element_wise
 } // namespace tensor_operation
 } // namespace ck
-#endif

include/ck/tensor_operation/gpu/element/element_wise_reduce_operation.hpp

@@ -0,0 +1,24 @@
+#pragma once
+#include "data_type.hpp"
+
+namespace ck {
+namespace tensor_operation {
+namespace element_wise {
+
+struct ReduceSum
+{
+    __host__ __device__ static constexpr float GetReduceZeroValue() { return float(0); }
+
+    __host__ __device__ void Reduce(float& acc, float v) const { acc += v; }
+};
+
+struct ReduceSquareSum
+{
+    __host__ __device__ static constexpr float GetReduceZeroValue() { return float(0); }
+
+    __host__ __device__ void Reduce(float& acc, float v) const { acc += v * v; }
+};
+
+} // namespace element_wise
+} // namespace tensor_operation
+} // namespace ck