[PATCH] D27692: [x86] use a single shufps when it can save instructions

Simon Pilgrim via Phabricator via llvm-commits llvm-commits at lists.llvm.org
Tue Dec 13 11:03:16 PST 2016


RKSimon added a comment.

In https://reviews.llvm.org/D27692#621231, @zansari wrote:

> Thanks for the link, Sanjay. Yes, I was just about to comment on this in the other review as i just got confirmation.. The info in that link is right. The h/w shufflers cross both domains after IVB, therefore, not suffering the bypass penalty when switching through such instructions (perm/shuf/unpack...).


So does that mean both this and https://reviews.llvm.org/D27684 can be committed safely? For recent hardware it makes no difference (and https://reviews.llvm.org/D27684 possibly saves a few instruction bytes). For older hardware we still save cycles compared to performing the extra shuffles and we should fix up the domain switches where possible to help a little more.


https://reviews.llvm.org/D27692





More information about the llvm-commits mailing list