[OpenMP] Dropped unnecessary define when compiling deviceRTLs for NVPTX

[tianshilei1992]

Summary

The comment said CUDA 9 header files use the nv_weak attribute which
clang is not yet prepared to handle. It's three years ago and now things have
changed. Based on my test, removing the definition doesn't have any problem on
my machine with CUDA 11.1 installed.

[jdoerfert]

Sure. But don't get too distracted here. All of the nvcc stuff will go away.

[tianshilei1992]

In D94700#2498764, @jdoerfert wrote:

Sure. But don't get too distracted here. All of the nvcc stuff will go away.

We still need some target specific code in any case. If we can eventually get rid of CUDA headers, then this could be useless. :-)

[jdoerfert]

In D94700#2498793, @tianshilei1992 wrote:

In D94700#2498764, @jdoerfert wrote:

Sure. But don't get too distracted here. All of the nvcc stuff will go away.

We still need some target specific code in any case. If we can eventually get rid of CUDA headers, then this could be useless. :-)

We don't for the device runtime. That is the cool thing ;)

openmp/libomptarget/deviceRTLs/nvptx/CMakeLists.txt

set(bc_flags ${LIBOMPTARGET_NVPTX_SELECTED_CUDA_COMPILER_FLAGS}
                  -I${devicertl_nvptx_directory}/src)
 
     if(${LIBOMPTARGET_NVPTX_DEBUG})
       set(bc_flags ${bc_flags} -DOMPTARGET_NVPTX_DEBUG=-1)
     else()
       set(bc_flags ${bc_flags} -DOMPTARGET_NVPTX_DEBUG=0)
     endif()
 
-    # CUDA 9 header files use the nv_weak attribute which clang is not yet prepared
-    # to handle. Therefore, we use 'weak' instead. We are compiling only for the
-    # device, so it should be equivalent.
-    if(CUDA_VERSION_MAJOR GREATER 8)
-      set(bc_flags ${bc_flags} -Dnv_weak=weak)
-    endif()
-
     # Create target to build all Bitcode libraries.
     add_custom_target(omptarget-nvptx-bc)
 
     # Generate a Bitcode library for all the compute capabilities the user requested.
     foreach(sm ${nvptx_sm_list})
       set(cuda_arch --cuda-gpu-arch=sm_${sm})
 
       # Compile CUDA files to bitcode.