AmpereOne Performance With GCC vs. LLVM Clang Compilers
Written by Michael Larabel in Software on 16 September 2024 at 09:49 AM EDT. Page 4 of 5. 10 Comments.
7-Zip compression performance was in favor of GCC on this AmpereOne A192-32X server but the decompression performance was faster under the Clang compiler builds.
AVIF image encoding was much faster when libavif was compiled by GCC but at least LLVM Clang 19/20 did show some nice speed-ups over Clang 18.
👁 C-Ray benchmark with settings of Resolution: 4K, Rays Per Pixel: 16. Clang 18.1.3 was the fastest.
👁 C-Ray benchmark with settings of Resolution: 5K, Rays Per Pixel: 16. Clang 18.1.3 was the fastest.
The C-Ray ray-tracer was fastest on LLVM Clang 18/19 but with seemingly a regression on Clang 20 Git. All the Clang versions though were faster than GCC 13 on Ubuntu 24.04 LTS ARM64.
GCC did turnaround a slight lead with the POV-Ray ray-tracer.