[llvm-branch-commits] [llvm] [LV] Choose best reduction for VPlan (PR #166138)

[Sam Tebbs via llvm-branch-commits]

================
@@ -0,0 +1,136 @@
+; NOTE: Assertions have been autogenerated by utils/update_test_checks.py UTC_ARGS: --version 6
+; RUN: opt -S -mcpu=neoverse-v2 -passes=loop-vectorize -mtriple=aarch64 < %s | FileCheck %s
+target triple = "aarch64"
+
+; Check that a partial reduction is reverted back to a regular reduction,
+; so that we compare "the VPlan with the best kind of reduction for <range>"
+; vs "the VPlan with the best kind of reduction for <other range>",
+
+; Function Attrs: nofree norecurse nosync nounwind memory(argmem: read) uwtable vscale_range(1,16)
+define dso_local i64 @foo(ptr noundef readonly captures(none) %0, i32 noundef %1) local_unnamed_addr #0 {
----------------
SamTebbs33 wrote:

Could you rename this test to something like `revert_add` and remove `local_unnamed_addr #0`?

https://github.com/llvm/llvm-project/pull/166138