[flang-commits] [flang] [flang] Extend LICM cheap nested hoisting to scalar loads (PR #217616)

[Slava Zakharin via flang-commits]

vzakhari wrote:

> I was wondering if there is a plan to enable this pass by default? Are there any current blockers on that?

I have been testing this with my custom pass pipeline for quite a while. The most recent issue was related to inferior compilation time - this was fixed by adding caching to FIR AA. The pass looks good to me as of today. We may find more issues once we enable it by default, but that is normal.

Feel free to flip the switch.

Please give me some time to test this patch downstream. Can you wait until Monday (8/24)?

https://github.com/llvm/llvm-project/pull/217616