AMD AOCC 3.0 Compiler Performance With The EPYC 75F3 - Making Fast Even Faster

Written by Michael Larabel in Software on 25 March 2021. Page 4 of 5. 2 Comments
EPYC 75F3 2P AOCC Compiler
EPYC 75F3 2P AOCC Compiler
EPYC 75F3 2P AOCC Compiler

The AMD AOCC compilers were also providing faster binaries for running the libavif AVIF image encoder compared to GCC on Ubuntu 20.04 LTS.

EPYC 75F3 2P AOCC Compiler
EPYC 75F3 2P AOCC Compiler
EPYC 75F3 2P AOCC Compiler
EPYC 75F3 2P AOCC Compiler
EPYC 75F3 2P AOCC Compiler
EPYC 75F3 2P AOCC Compiler
EPYC 75F3 2P AOCC Compiler
EPYC 75F3 2P AOCC Compiler
EPYC 75F3 2P AOCC Compiler
EPYC 75F3 2P AOCC Compiler

AOCC also really paid off when being used to compile Intel's oneDNN deep learning library. The AOCC-built oneDNN tended to perform measurably better than GCC, which will be interesting to see then how these results compare to current Intel Xeon processors with that being one of the areas where Cascade Lake still competed well with Milan was in Intel's own oneDNN open-source library.


Related Articles