<div dir="ltr"><div class="gmail_extra"><br><div class="gmail_quote">On Wed, Jan 15, 2014 at 4:13 PM, Diego Novillo <span dir="ltr"><<a href="mailto:dnovillo@google.com" target="_blank">dnovillo@google.com</a>></span> wrote:<br>
<blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex"><div id=":bla" style="overflow:hidden">Chandler also pointed me at the vectorizer, which has its own<br>
unroller. However, the vectorizer only unrolls enough to serve the<br>
target, it's not as general as the runtime-triggered unroller. From<br>
what I've seen, it will get a maximum unroll factor of 2 on x86 (4 on<br>
avx targets). Additionally, the vectorizer only unrolls to aid<br>
reduction variables. When I forced the vectorizer to unroll these<br>
loops, the performance effects were nil.</div></blockquote></div><br>I just also want to point out that we should really be *vectorizing* this loop as well. It's a great candidate for it AFAICS....</div></div>