[PATCH] D32104: Teach SCEV normalization to de/normalize non-affine add recs

[Sanjoy Das via Phabricator via llvm-commits]

sanjoy added a comment.

In https://reviews.llvm.org/D32104#733847, @atrick wrote:

> Neat. I wouldn't usually add complexity for an unexpected situation, but this is purely a generalization.


Yes, it does not really matter from a feature perspective, but it means we no longer need to be careful to filter out specific add recurrences from IVUsers et. al..  Speaking of which, can you please also take a look at https://reviews.llvm.org/D32105?


https://reviews.llvm.org/D32104