GCC & LLVM Clang Compiler Benchmarks On AMD's EPYC 7601
When moving from GCC to LLVM Clang, the performance was up to 6% faster than GCC.
For programs like x264 where there is already fine-tuning in Assembly by its developers for maximum performance, the compiler doesn't get in the way much.
While LLVM Clang supports OpenMP multi-threading, in most benchmarks it's still not as efficient as GCC. With a high core count rig like with AMD EPYC, OpenMP tends to be very important.
If concerned about the time needed to compile software, LLVM Clang still generally is faster at compiling code than GCC. But Clang's fast build speeds have resulted in GCC developers making improvements there too in recent releases.
GCC meanwhile produces faster binaries for testing C-Ray multi-threaded ray-tracing.