[llvm] [NFCI][VPlan] Split initial mem-widening into a separate transformation (PR #182592)

[Florian Hahn via llvm-commits]

================
@@ -0,0 +1,114 @@
+; NOTE: Assertions have been autogenerated by utils/update_test_checks.py UTC_ARGS: --check-globals none --version 6
----------------
fhahn wrote:

for the name, may be good to call it `ordered-reduction-with-invariant-stores.ll`, similar to `partial-reduce-with-invariant-stores.ll` in same dir.

Not sure if this relies on the ARM64, but there is an option to force ordered reductions (-force-ordered-reductions), and with that it may be possible to make target independent, not sure.

https://github.com/llvm/llvm-project/pull/182592