[PING] [PATCH] faster string operations for bulldozer (take 2)

[Roland McGrath]

> On bulldozer unaligned loads have small penalty. You could read about
> buldozer improvements or test it by simple benchmark.
> 
> And what evidence that avx on intel processors is always
> correlated with efficiency of unaligned loads?

I don't understand what you're saying or how it constitutes a sensical
answer to my question.  If a particular model or family of chip models has
efficient unaligned loads, then setting the flag should be contingent on
that model/family detection, not on AVX support.  Your change makes it
contingent only on AVX support, with no explanation of why that should be
taken as an indicator of efficient unaligned loads.