batched_gemm: use profiler in ctest (#163)

include/ck/tensor_operation/gpu/device/device_gemm.hpp

@@ -1,5 +1,7 @@
 #pragma once
 #include <iostream>
+#include <vector>
+
 #include "device_base.hpp"
 
 namespace ck {

include/ck/tensor_operation/gpu/device/tensor_layout.hpp

@@ -1,5 +1,4 @@
-#ifndef TENSOR_LAYOUT_HPP
-#define TENSOR_LAYOUT_HPP
+#pragma once
 
 namespace ck {
 namespace tensor_layout {
@@ -128,4 +127,3 @@ std::ostream& operator<<(std::ostream& os, const Layout&)
 
 } // namespace tensor_layout
 } // namespace ck
-#endif