add --offload-compress compiler flag (#1433)

* add --offload-compress compiler flag

* only apply the --offload-compress flag to the ckProfiler

* move the --offload-compress flag back to main cmake file

* add offload-compress to target compile option of ckProfiler

---------
Co-authored-by: carlushuang <carlus.huang@amd.com>

library/src/tensor_operation_instance/gpu/CMakeLists.txt

@@ -87,6 +87,13 @@ function(add_instance_library INSTANCE_NAME)
         endforeach()
         add_library(${INSTANCE_NAME} OBJECT ${INST_OBJ})
         target_compile_features(${INSTANCE_NAME} PUBLIC)
+
+        # flags to compress the library
+        if(NOT WIN32 AND ${hip_VERSION_FLAT} GREATER 600241132)
+            message("Adding --offload-compress flag for ${INSTANCE_NAME}")
+            target_compile_options(${INSTANCE_NAME} PRIVATE --offload-compress)
+        endif()
+
         set_target_properties(${INSTANCE_NAME} PROPERTIES POSITION_INDEPENDENT_CODE ON)
         clang_tidy_check(${INSTANCE_NAME})
         set(result 0)

profiler/src/CMakeLists.txt

@@ -82,6 +82,11 @@ set(PROFILER_EXECUTABLE ckProfiler)
 
 add_executable(${PROFILER_EXECUTABLE} ${PROFILER_SOURCES})
 target_compile_options(${PROFILER_EXECUTABLE} PRIVATE -Wno-global-constructors)
+# flags to compress the library
+if(NOT WIN32 AND ${hip_VERSION_FLAT} GREATER 600241132)
+  message("Adding --offload-compress flag for ${PROFILER_EXECUTABLE}")
+  target_compile_options(${PROFILER_EXECUTABLE} PRIVATE --offload-compress)
+endif()
 
 target_link_libraries(${PROFILER_EXECUTABLE} PRIVATE utility getopt::getopt)
 target_link_libraries(${PROFILER_EXECUTABLE} PRIVATE device_gemm_instance)