GCC 8/9 vs. LLVM Clang 7/8 Compiler Performance On POWER9 With The Raptor Talos II
Like on x86_64, building with GCC yields faster performance for C-Ray.
GCC also yielded faster IBM POWER binaries for the AOBench ambient occlusion renderer compared to Clang.
GCC also assembled faster binaries for the Bullet physics engine compared to Clang.
GCC also yielded faster binaries for the dav1d AV1 video decoder, including a nice performance speed-up in moving from GCC8 to GCC9.
FLAC and LAME MP3 audio encoding also performed better when built by GCC, especially with the upcoming GCC9 compiler release.