LLVM/Clang 3.2 Compiler Competing With GCC
Written by Michael Larabel in Software on 27 December 2012. Page 3 of 5. 20 Comments

The Sparse Matrix Multiply test is noticeably faster with LLVM/Clang 3.2 over the earlier 3.1 release, which secures a nice lead over GCC 4.7 and the 4.8 development code.

The LU Matrix Factorization benchmark in SciMark regresses from LLVM/Clang 3.1 to LLVM/Clang 3.2 that puts its performance back in line with GCC.

LLVM/Clang 3.2 continues to lead over the Free Software Foundation's compiler with Jacobi Successive Over-Relaxation.

LLVM/Clang continues to struggle with the John The Ripper benchmark since it still lacks OpenMP support.