> > So, is it OK to commit this patch to trunk, and H.J's original patch + this > to 4.6 branch?
>I have no problems on -mtune=Bulldozer. But I object -mtune=generic >change and did suggest a different approach for -mtune=generic. What's your suggested approach for -mtune=generic? My understanding of putting something into -mtune=generic is that all are happy, or some are willing to lose. This rule guided us not to turn on software prefetching for -mtune=generic. I would like to hear your suggestion that sandy bridge could get the performance while bulldozer does not lose from splitting avx256 unaligned loads! Thanks, Changpeng