[llvm] [AArch64] Improve codegen of vectorised early exit loops (PR #119534)

[Florian Hahn via llvm-commits]

================
@@ -0,0 +1,195 @@
+; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
+; RUN: llc < %s -mtriple=aarch64-none-linux-gnu -mattr=+sve | FileCheck %s
----------------
fhahn wrote:

The transform is done on IR in CodeGenPrepare, right? It would probably be easier to see what is going on exactly if there would be at least some IR tests in `llvm/test/Transforms/CodeGenPrepare/AArch64`?

https://github.com/llvm/llvm-project/pull/119534