[llvm] [LoopVectorizer] Add support for chaining partial reductions (PR #120272)

[Nicholas Guy via llvm-commits]

================
@@ -0,0 +1,568 @@
+; NOTE: Assertions have been autogenerated by utils/update_test_checks.py UTC_ARGS: --version 4
+; RUN: opt -S -o - %s --passes=loop-vectorize -force-vector-width=16 -force-vector-interleave=1 -scalable-vectorization=on | FileCheck %s
----------------
NickGuy-Arm wrote:

It was needed as otherwise `-force-vector-width=16` disables scalable vectorization. I've switched it to use `-vectorizer-maximize-bandwidth` now though, which achieves the same thing.

https://github.com/llvm/llvm-project/pull/120272