[llvm] [X86][LoopVectorize] Enable MaximizeBandwidth by default on X86 with generic split cost floor (PR #201666)

[Sumukh J Bharadwaj via llvm-commits]

================
@@ -1,6 +1,6 @@
 ; NOTE: Assertions have been autogenerated by utils/update_test_checks.py UTC_ARGS: --check-globals none --version 6
-; RUN: opt -passes=loop-vectorize -S < %s 2>&1 | FileCheck %s --check-prefix=X86
-; RUN: opt -passes=loop-vectorize -mattr=+avx512f -S < %s 2>&1 | FileCheck %s --check-prefix=AVX512
+; RUN: opt -passes=loop-vectorize -vectorizer-maximize-bandwidth=false -S < %s 2>&1 | FileCheck %s --check-prefix=X86
+; RUN: opt -passes=loop-vectorize -mattr=+avx512f -vectorizer-maximize-bandwidth=false -S < %s 2>&1 | FileCheck %s --check-prefix=AVX512
----------------
amd-subharad wrote:

Agreed, this is not ideal. These tests were not designed to test MaxBW behavior, so we added the flag to preserve their original intent while the default changes. We will regenerate the CHECK lines with MaxBW enabled so they validate the new default behavior.

https://github.com/llvm/llvm-project/pull/201666