GCC 8/9 vs. LLVM Clang 6/7 Compiler Benchmarks On AMD EPYC
With the Stockfish chess engine benchmark was one of the rare scenarios where LLVM Clang 7.0 SVN at "-O3 -march=native" was delivering slightly faster performance over GCC 8/9.
The Redis performance results were mixed between GCC and Clang at some of the optimization levels tested.
The GCC-built PostgreSQL binaries were generally delivering a slight lead over Clang but in some instances the performance was effectively identical.
Clang was delivering a lead over GCC at the -O2 optimization level but was tied at -O3 and GCC pulled ahead when it came to the -march=native running on this AMD EPYC processor.