On 01/04/14 17:02, John Steele Scott wrote: > > GCC 4.8 added an SSE2 optimized simd_fast_mersenne_twister_engine in the > __gnu_cxx namespace. I've been able to use this successfully with both GCC > and > ICC, but when I use it with Clang, it doesn't give consistent results.
Ah, this is <http://llvm.org/bugs/show_bug.cgi?id=18323>. There's a workaround there which fixes my test case as well. Cheers, John _______________________________________________ cfe-users mailing list cfe-users@cs.uiuc.edu http://lists.cs.uiuc.edu/mailman/listinfo/cfe-users
