[OpenMP][NVPTX] Added the missing -O1 when building NVPTX bitcode libraries

[tianshilei1992]

Summary

In the past -O1 was used when building NVPTX bitcode libraries. After
we switched to OpenMP, -O1 was missing by mistake, leading to a huge performance
regression.

[tianshilei1992]

openmp/libomptarget/deviceRTLs/nvptx/CMakeLists.txt

121

Also align these commands at different lines

[JonChesterfield]

Ah, yes. That would be slower. Thanks

[jdoerfert]

Does O1 create TBAA? we might need -O3 -Xclang -disable-llvm-optzns or -Xclang -disable-O0-optnone if we can also remove the noinline that way.

[jdoerfert]

we need to put this in 12 as well, I think we need to open a bug that blocks the 12 meta-bug and start adding these patches (or git hashes) to be included in the release branch.

[tianshilei1992]

In D95545#2525820, @jdoerfert wrote:

Does O1 create TBAA? we might need -O3 -Xclang -disable-llvm-optzns or -Xclang -disable-O0-optnone if we can also remove the noinline that way.

Yes, TBAA is created.

[tianshilei1992]

rebase

[tianshilei1992]

rebased

openmp/libomptarget/deviceRTLs/nvptx/CMakeLists.txt

set(cuda_src_files
   ${devicertl_common_directory}/src/reduction.cu
   ${devicertl_common_directory}/src/support.cu
   ${devicertl_common_directory}/src/sync.cu
   ${devicertl_common_directory}/src/task.cu
   src/target_impl.cu
 )
 
 # Set flags for LLVM Bitcode compilation.
-set(bc_flags -S -x c++
-              -target nvptx64
-              -Xclang -emit-llvm-bc
-              -Xclang -aux-triple -Xclang ${aux_triple}
-              -fopenmp -fopenmp-cuda-mode -Xclang -fopenmp-is-device
-              -D__CUDACC__
-              -I${devicertl_base_directory}
-              -I${devicertl_nvptx_directory}/src)
+set(bc_flags -S -x c++ -O1 -std=c++14
 
 if(${LIBOMPTARGET_NVPTX_DEBUG})
   list(APPEND bc_flags -DOMPTARGET_NVPTX_DEBUG=-1)
 else()
   list(APPEND bc_flags -DOMPTARGET_NVPTX_DEBUG=0)
 endif()
 
 # Create target to build all Bitcode libraries.