GEMM with Multiple Source, GEMM+Bias+Add+FastGeLU example and ckProfiler (#241)

* ad gelu and fast_gelu

* added GeLU and fast GeLU

* clean up

* add gemm+fastgelu example

* add gemm+gelu instances

* update profiler

* clean up

* clean up

* adding gemm+bias+activation

* clean

* adding bias

* clean

* adding gemm multiple d

* debugging

* add gemm bias add fastgelu

* rename, clean

* refactoring; add readme

* refactor

* refactor

* refactor

* refactor

* refactor

* refactor

* fix

* fix

* update example

* update example

* rename

* update example

* add ckProfiler

* clean

* clean

* clean

* clean

* add comment

* use type_convert

* clean

* clean element wise op

[ROCm/composable_kernel commit: 56adf7e9cc]

library/include/ck/library/reference_tensor_operation/cpu/reference_gemm_bias_2d.hpp

@@ -66,8 +66,8 @@ struct ReferenceGemmBias2D : public device::BaseOperator
 
                 for(int k = 0; k < K; ++k)
                 {
-                    arg.a_element_op_(a, static_cast<AccDataType>(arg.a_m_k_(m, k)));
-                    arg.b_element_op_(b, static_cast<AccDataType>(arg.b_k_n_(k, n)));
+                    arg.a_element_op_(a, ck::type_convert<AccDataType>(arg.a_m_k_(m, k)));
+                    arg.b_element_op_(b, ck::type_convert<AccDataType>(arg.b_k_n_(k, n)));
                     acc += a * b;
                 }
 

library/include/ck/library/tensor_operation_instance/device_operation_instance.hpp

@@ -1,7 +1,6 @@
-#ifndef CK_DEVICE_OPERATION_INSTANCE_HPP
-#define CK_DEVICE_OPERATION_INSTANCE_HPP
+#pragma once
 
-#include <stdlib.h>
+#include <vector>
 
 namespace ck {
 namespace tensor_operation {
@@ -23,4 +22,3 @@ void add_device_operation_instances(std::vector<std::unique_ptr<OpInstance>>& op
 } // namespace device
 } // namespace tensor_operation
 } // namespace ck
-#endif