[PATCH] D13259: LLE 6/6: Add LoopLoadElimination pass

[Adam Nemet via llvm-commits]

anemet added a comment.

In http://reviews.llvm.org/D13259#279193, @hfinkel wrote:

> LGTM too, except for:
>
> I don't think we check anywhere to disqualify volatile/atomic loads/stores, and we need to do that (especially for volatiles, since you're changing the number of accesses, and I'd need to think more-carefully about atomics).


LAA stops the analysis unless the memops are all isSimple().  Let me add a comment about this.


http://reviews.llvm.org/D13259