[llvm] [LAA] Support different strides & non constant dep distances using SCEV. (PR #88039)

[Florian Hahn via llvm-commits]

fhahn wrote:

@psteinfeld thanks for the heads up. Unfortunately I don't think I have access to that benchmark and I am also not really familiar with flang. Could you share the IR with and without the patch? It is likely that the patch enabled additional vectorization.

https://github.com/llvm/llvm-project/pull/88039