Page MenuHomePhabricator

[SVE] Lower fixed length VECREDUCE_[UMAX|UMIN] to Scalable
ClosedPublic

Authored by cameron.mcinally on Sep 25 2020, 9:17 AM.

Details

Summary

Essentially the same as the signed variants from D88259. Also includes an attempt to clean up the lowering function.

Diff Detail

Event Timeline

cameron.mcinally requested review of this revision.Sep 25 2020, 9:17 AM
llvm/test/CodeGen/AArch64/sve-fixed-length-int-reduce.ll
1042

Just noticed this. Will correct...

Correct bad string replacement.

paulwalker-arm accepted this revision.Sep 28 2020, 4:28 AM
paulwalker-arm added a subscriber: kmclaughlin.

Just a heads up that @kmclaughlin is starting to look at legalisation/lowering for scalable vector types. From an operation legalisation point of view you've done most of the plumbing so hopefully any toe treading will be minimal.

This revision is now accepted and ready to land.Sep 28 2020, 4:28 AM

Just a heads up that @kmclaughlin is starting to look at legalisation/lowering for scalable vector types. From an operation legalisation point of view you've done most of the plumbing so hopefully any toe treading will be minimal.

Ok, no problem. I still have to finish the fixed FP reductions. If I get in your way, just shout.

Or, if @kmclaughlin would like, I can add the Scalable lowerings while I'm here. Assuming you don't have downstream changes prepared already...

Or, if @kmclaughlin would like, I can add the Scalable lowerings while I'm here. Assuming you don't have downstream changes prepared already...

There's the possibility of some type legalisation changes as well, so best to leave the scalable lowerings for @kmclaughlin.