merge develop

composable_kernel/include/utility/common_header.hpp

@@ -35,8 +35,8 @@
 #include "dynamic_buffer.hpp"
 #include "is_known_at_compile_time.hpp"
 #include "transpose_vectors.hpp"
-
 #include "inner_product.hpp"
+#include "element_wise_operation.hpp"
 
 // TODO: remove this
 #if CK_USE_AMD_INLINE_ASM

composable_kernel/include/utility/utility.hpp

@@ -5,8 +5,12 @@
 
 namespace ck {
 
+__device__ constexpr index_t get_wave_size() { return CK_GPU_WAVE_SIZE; }
+
 __device__ index_t get_thread_local_1d_id() { return threadIdx.x; }
 
+__device__ index_t get_wave_local_1d_id() { return threadIdx.x / get_wave_size(); }
+
 __device__ index_t get_block_1d_id() { return blockIdx.x; }
 
 } // namespace ck