An Early Look At LLVM Clang 13 Performance On AMD Zen 3

Written by Michael Larabel in Software on 12 August 2021 at 02:47 PM EDT. Page 1 of 4. 7 Comments.

With LLVM/Clang 13 feature development having ended last week and the 13.0 release candidate being tagged, in starting off the benchmarking cycle first up I was looking at how well this new compiler is performing compared to LLVM Clang 12 stable on an AMD EPYC 7543 (Zen 3) Linux server.

LLVM 13 is shaping up nicely as the H2'2021 update to this leading and widely-used open-source compiler stack. Early benchmarks I've conducted so far on the branched LLVM 13 code is looking positive as an incremental improvement over LLVM Clang 12 that is already very competitive to GCC on x86_64 and AArch64 hardware.

LLVM 13.0 stable isn't expected until late September with at least two more release candidates due out before then. Given it's just been over one week since the code branching, I just ran some preliminary benchmarks so far on an AMD EPYC 7003 series server and will have more compiler benchmarks and on more hardware when the actual September release approaches.

LLVM Clang 13 Early Benchmarks On AMD EPYC Zen 3

Given the popularity of AMD EPYC Milan and Zen 3 at large, this round of initial benchmarking was performed on an AMD EPYC 7543 1P server -- the Tyan Transport CX GC68B8036-LE. Clang 12.0.1 against a Clang 13.0 daily snapshot from 9 August was used for benchmarking with both compiler builds obtained via apt.llvm.org.

Related Articles