LLVM/Clang 3.3 Performing Against GCC For Old Intel CPU

Written by Michael Larabel in Software on 23 April 2013. Page 2 of 3. 5 Comments

GCC 4.7.3 had a slight advantage over LLVM/Clang when it came to running the HMMer scientific benchmark on the Intel Core 2 Duo T5600 Mac Mini.

The MAFFT compiler performance was nearly matched between GCC and LLVM's Clang.

One of the shortcomings right now for LLVM/Clang is lacking OpenMP multi-processing support, but with a dual-core CPU this is less of an impact than a modern 8+ core system. Even without OpenMP support, LLVM/Clang 3.2 and 3.3 SVN ended up being faster than GCC 4.7.3 in GraphicsMagick. For many operations with this open-source imaging application, the LLVM/Clang performance was noticeably better than the GNU Compiler Collection.

Related Articles