[llvm] [LV] Linear argument tests for vectorization of function calls (PR #73936)

[Maciej Gabka via llvm-commits]

================
@@ -0,0 +1,275 @@
+; NOTE: Assertions have been autogenerated by utils/update_test_checks.py UTC_ARGS: --version 2
+; RUN: opt < %s -passes=loop-vectorize,instsimplify -force-vector-interleave=1 -S | FileCheck %s
----------------
mgabka wrote:

Hi @huntergr-arm ,
wouldn't be better to remove the "test_linear_with_mask" test function, but:
* have more than 1 mapping available per scalar call, (as that is most likely going to be real life scenario), so we can test if LV is choosing the right one,
* add more RUN lines?
1. RUN: opt < %s -passes=loop-vectorize,instsimplify -force-vector-interleave=1 //only Neon mappings used
2.RUN: opt < %s -passes=loop-vectorize,instsimplify -force-vector-interleave=1 -mattr=+sve //only sve mappings used
3.RUN: opt < %s -passes=loop-vectorize,instsimplify -force-vector-interleave=1 -mattr=+sve -prefer-predicate-over-epilogue=predicate-dont-vectorize //sve mappings used, but also Neon masked mapping used when available

https://github.com/llvm/llvm-project/pull/73936