search for: scalarrevolut

...[2*k] = x[2*k]+y[k]; } } Vectorization failed due to "unsafe dependent memory operation". I traced the LoopAccessAnalysis.cpp and found the reason is the NoWrapFlag for SCEVAddRecExpr is not set and consequently the dependence distant became unknown. Can anyone familiar with ScalarRevolution tell me whether this is an expected behavior or a bug? Tong -------------- next part -------------- An HTML attachment was scrubbed... URL: <http://lists.llvm.org/pipermail/llvm-dev/attachments/20150610/7949fbac/attachment.html>

[+Arnold] > On Jun 10, 2015, at 1:29 PM, Sanjoy Das <sanjoy at playingwithpointers.com> wrote: > > [+CC Andy] > >> Can anyone familiar with ScalarRevolution tell me whether this is an >> expected behavior or a bug? > > Assuming you're talking about 2*k, this is a bug. ScalarEvolution > should be able to prove that {0,+,4} is <nsw> and <nuw>. I also find it surprising that the inbounds gep does not allow us to prove...

...ed, Jun 10, 2015 at 5:29 PM, Adam Nemet <anemet at apple.com> wrote: >> [+Arnold] >> >>> On Jun 10, 2015, at 1:29 PM, Sanjoy Das <sanjoy at playingwithpointers.com> wrote: >>> >>> [+CC Andy] >>> >>>> Can anyone familiar with ScalarRevolution tell me whether this is an >>>> expected behavior or a bug? >>> >>> Assuming you're talking about 2*k, this is a bug. ScalarEvolution >>> should be able to prove that {0,+,4} is <nsw> and <nuw>. >> >> I also find it surprising...