FreeBSD 9.1: LLVM/Clang Battling GCC
LLVM Clang carried a modest performance against over GCC for Himeno on the Intel Core i7 system
There were no meaningful results out of 7-Zip.
LLVM Clang 3.1 had a small performance advantage over GCC 4.2.1 for the C-Ray multi-threaded ray-tracing.
GCC won with Primesieve and N-Queens again because of LLVM's lack of OpenMP support.
Overall, LLVM/Clang 3.1 as found in FreeBSD/PC-BSD 9.1 was generally faster than GCC 4.2.1, the last release that was licensed under the GNU GPLv2. Clang being faster though isn't a big surprise seeing as version 3.1 was released in 2012 while GCC 4.2.1 dates back to 2007. The areas where the GNU Compiler Collection managed to produce faster binaries on FreeBSD were in the OpenMP programs where LLVM/Clang still lacks support for the Open MultiProcessing API.
If you enjoyed this article consider joining Phoronix Premium to view this site ad-free, multi-page articles on a single page, and other benefits. PayPal or Stripe tips are also graciously accepted. Thanks for your support.
