Florian, > * H. J. Lu: > > > Here are diffs of SPEC CPU 2K between before and after with gcc 4.1 > > using "-O2 -ffast-math" on Nocona: > > And what about Opterons? IOW, how "generic" is the optimization?
The generic code generation should cost a small compromise in performance relative to native code generation (e.g., -mtune=opteron on Opteron), but without affecting performance significantly on other processors (e.g., -mtune=nocona on Opteron). The goal of -mtune=generic is that it's better than the latter case albeit slightly slower than the former case. And, yes, AMD co-developped the -mtune=generic patch. HTH -- _______________________________________________________ Evandro Menezes AMD Austin, TX
