GCC 4.8.0 vs. LLVM Clang 3.3 Compiler Performance Three-Way
Smallpt is just another case where LLVM/Clang performs slowly due to its lack of OpenMP support.
The FFmpeg performance hasn't really changed much with the new LLVM/Clang release, but this video encoder relies a lot on hand-tuned Assembly.
The PostgreSQL server performance also hasn't evolved much with LLVM/Clang 3.3.
For the Intel systems with the Apache web-server the LLVM/Clang 3.3 performance appears slightly faster.
Overall, LLVM/Clang 3.3 appears to be largely competitive with GCC 4.8.0 except in select cases like the missing OpenMP support. Depending upon the hardware, there are some nice performance improvements with LLVM 3.3 over its predecessor. Aside from performance changes, LLVM 3.3 is nice in that the Clang compiler is C++11 feature-complete, there is now the ARM AArch64 64-bit support, IBM SystemZ support, there's the SLP Vectorizer and improvements to the Loop Vectorizer, the AMD R600 GPU LLVM back-end has been merged, and much more. Look for the LLVM/Clang 3.3 final release in early June.
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.