[flang-commits] [flang] [flang][NFC] Engineering option for optimizing allocatable assignments. (PR #119936)

[Slava Zakharin via flang-commits]

vzakhari wrote:

I ran some SPEC and Polyhedron benchmarks on x86, and did not find any noticeable improvements from enabling this option. Of course, the ability to optimize an assignment with allocatable LHS depends on the alias analysis. So we may not see any benefits now, but maybe we will see some later. I would like to keep this option, so that we can quickly run an experiment for an app where we think this might be profitable.

https://github.com/llvm/llvm-project/pull/119936