Ryzen Compiler Performance: Clang 4/5 vs. GCC 6/7/8 Benchmarks
GCC on Ryzen with just -march=x86-64 is noticeably faster than Clang in a similar configuration while the znver1 optimization level does regress the performance.
GCC with znver1 is faster than Clang for LAME MP3 encoding.
While the WavPack audio encoding saw no real changes across the tested compilers.
libjpeg-turbo's tjbench benchmark hadn't encountered any real changes over the tested compilers.