Warnings with nvc++ as the compiler

[crtrott]

Describe the bug

The following warnings were observed when building googletest:

[ 12%] Building CXX object googletest/CMakeFiles/gtest.dir/src/gtest-all.cc.o
"/ascldap/users/crtrott/Software/googletest/googletest/src/gtest-internal-inl.h", line 636: warning: unknown attribute "optimize"
        GTEST_NO_INLINE_ GTEST_NO_TAIL_CALL_;
                         ^

"/ascldap/users/crtrott/Software/googletest/googletest/src/gtest.cc", line 6252: warning: unknown attribute "optimize"
  GTEST_NO_INLINE_ GTEST_NO_TAIL_CALL_ std::string
                   ^

[ 25%] Linking CXX static library ../lib/libgtest.a
[ 25%] Built target gtest
Scanning dependencies of target gmock
Scanning dependencies of target gtest_main
[ 37%] Building CXX object googletest/CMakeFiles/gtest_main.dir/src/gtest_main.cc.o
[ 50%] Building CXX object googlemock/CMakeFiles/gmock.dir/src/gmock-all.cc.o
"/ascldap/users/crtrott/Software/googletest/googlemock/src/gmock-spec-builders.cc", line 674: warning: variable "l" was declared but never referenced
    internal::MutexLock l(&internal::g_gmock_mutex);

Steps to reproduce the bug

cmake -DCMAKE_CXX_STANDARD=17 -DCMAKE_CXX_COMPILER=nvc++ ../
-- The C compiler identification is GNU 4.8.5
-- The CXX compiler identification is NVHPC 22.3.0
-- Detecting C compiler ABI info
-- Detecting C compiler ABI info - done
-- Check for working C compiler: /usr/bin/cc - skipped
-- Detecting C compile features
-- Detecting C compile features - done
-- Detecting CXX compiler ABI info
-- Detecting CXX compiler ABI info - done
-- Check for working CXX compiler: /ascldap/users/projects/x86-64/nvidia/hpc_sdk/Linux_x86_64/22.3/compilers/bin/nvc++ - skipped
-- Detecting CXX compile features
-- Detecting CXX compile features - done
-- Found Python: /usr/bin/python3.6 (found version "3.6.8") found components: Interpreter 
-- Looking for pthread.h
-- Looking for pthread.h - found
-- Performing Test CMAKE_HAVE_LIBC_PTHREAD
-- Performing Test CMAKE_HAVE_LIBC_PTHREAD - Failed
-- Looking for pthread_create in pthreads
-- Looking for pthread_create in pthreads - not found
-- Looking for pthread_create in pthread
-- Looking for pthread_create in pthread - found
-- Found Threads: TRUE  
-- Configuring done
-- Generating done
-- Build files have been written to: /ascldap/users/crtrott/Software/googletest/build

Does the bug persist in the most recent commit?

This was tested with the latest commit 8d51dc5

What operating system and version are you using?

RedHat 7.x

Linux kokkos-dev-2 3.10.0-1160.36.2.el7.x86_64 #1 SMP Thu Jul 8 02:53:40 UTC 2021 x86_64 x86_64 x86_64 GNU/Linux

What compiler and version are you using?

nvc++ --version

nvc++ 22.3-0 64-bit target on x86-64 Linux -tp skylake-avx512 
NVIDIA Compilers and Tools
Copyright (c) 2022, NVIDIA CORPORATION & AFFILIATES.  All rights reserved.

What build system are you using?

cmake version 3.21.1

Additional context

The first issue can be resolved by checking for __NVCOMPILER in

googletest/googletest/include/gtest/internal/gtest-port.h

Lines 772 to 779 in 8d51dc5

	#define GTEST_NO_TAIL_CALL_ __attribute__((disable_tail_calls))
	#endif
	#elif __GNUC__
	#define GTEST_NO_TAIL_CALL_ \
	__attribute__((optimize("no-optimize-sibling-calls")))
	#else
	#define GTEST_NO_TAIL_CALL_
	#endif

I.e. the rootcause seems to be that __NVCOMPILER does define GNUC, but its not really GCC. So just making sure that the GCC variant is only used when __NVCOMPILER is not defined fixes that warning.

I would issue a PR, but not sure its worth the effort to get the corporate contributor agreement signed ..

The second warning, is likely on nvc++ itself (i mean the lock is clearly used and serves a function), however one could fix it with adding a (void) l in the code.

[crtrott]

@brycelelbach @mhoemmen not sure if NVIDIA has a contributor agreement, and you guys could issue a PR?

[dinord]

We'd be willing to accept a pull request to fix this issue. We do not plan to implement a fix ourselves, since we do not officially support nvc++.

[mhoemmen]

@crtrott I'm new to NVIDIA, so I'm asking about this.