[PATCH] D28676: Makes incremental dominator calculation in Loop Unroll pass

[Eli Friedman via Phabricator via llvm-commits]

efriedma added a comment.

Making a quick comparison between the two patches, the changes to LoopUnrollRuntime.cpp look almost identical.  I think my version of the changes to the dominance calculation in LoopUnroll.cpp are more general.  And obviously my changes to peeling don't actually dynamically update the domtree.  I should probably merge the two loops in CloneLoopBlocks like this patch does.

Serge, do you have any other comments on my patch?


https://reviews.llvm.org/D28676