Yes, I agree as well.
Nov 2 2017
I agree that this a strict improvement.
Oct 11 2017
It sounds like this is a profitability issue and therefore the LAA users should handle this? Maybe the users would also want to be more general (I think a loop with an iteration count of 2 is probably not worth vectorizing).
Sep 2 2015
Thanks for the reduced testcase!
Aug 2 2014
Jul 31 2014
Thanks, this all makes sense to me. However, I think there is still an issue with the default
Jul 30 2014
Jul 27 2014
Looks like I spoke too soon (new comment inlined). I think there are some issues with targets without a machine model.
Jul 25 2014
Ah, I just noticed that a header file is missing (AArch64MachineCombinerPattern.h).
Jul 18 2014
Just some minor issues that I've noticed (comments inlined). Otherwise looks good.
Jul 10 2014
It seems to me that the branch mispredict cost for the case where the values of X are random would outweigh the benefits of this transformation for your alternative code sequence, even on OoO cores.
Jul 9 2014
Jul 6 2014