AMD AOCC 3.2 Helps Squeeze A Bit More Out Of Zen 3

Overall, AOCC 3.2 seems to be a nice evolutionary update from the testing I have carried out thus far on EPYC Milan. In many cases the difference was quite small but did not encounter any notable regressions with the workloads tested and in a number of instances there were healthy improvements out of this incremental AOCC update.

62 different tests were carried out and overall AOCC 3.2 is another step in the right direction for AMD's compiler stack while it would be nice if AOCC sources were public to help track the delta over upstream LLVM/Clang. With the 62 tests carried out, the AOCC 3.2 built binary was the fastest 77% of the time. AOCC 3.2 was the slowest just in 2 of the 62 tests. See all of the numbers over on OpenBenchmarking.org.

On the holiday benchmarking agenda is also a fresh comparison looking at AOCC 3.2 performance against upstream LLVM/Clang and GCC along with the possible Ryzen comparison.