[llvm] [X86][CostModel] Price predicate mask-expansion fanout for wide predicated ops (PR #201666)

[Sumukh J Bharadwaj via llvm-commits]

================
@@ -0,0 +1,63 @@
+; NOTE: Assertions have been autogenerated by utils/update_analyze_test_checks.py UTC_ARGS: --version 6
+; RUN: opt < %s -passes="print<cost-model>" 2>&1 -disable-output -cost-kind=throughput -mtriple=x86_64-- -mattr=+sse2 | FileCheck %s -check-prefixes=CHECK,SSE
+; RUN: opt < %s -passes="print<cost-model>" 2>&1 -disable-output -cost-kind=throughput -mtriple=x86_64-- -mattr=+avx2 | FileCheck %s -check-prefixes=CHECK,AVX2
+; RUN: opt < %s -passes="print<cost-model>" 2>&1 -disable-output -cost-kind=throughput -mtriple=x86_64-- -mattr=+avx512f | FileCheck %s -check-prefixes=CHECK,AVX512
----------------
amd-subharad wrote:

Added the wide select cases to select.ll, whose RUN lines all use -cost-kind=all, so RThru/CodeSize/Lat/SizeLat are each checked
BLENDV is covered by its existing +sse4.1 run. I kept the -mattr tiers rather than x86-64-v{2,3,4} since they cover the same ISA levels and are finer. Would that be fine?

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