[llvm] [CodeGen] Fold load into LZCNT/TZCNT/POPCNT after compare elimination in peephole optimizer (PR #194662)

[Simon Pilgrim via llvm-commits]

================
@@ -0,0 +1,129 @@
+; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py UTC_ARGS: --version 4
+; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+lzcnt,+bmi,+popcnt | FileCheck %s
+
+; Test that loads are folded into the memory form of LZCNT/TZCNT/POPCNT when
----------------
RKSimon wrote:

rename this bitcnt-load-with-cmov.ll - to reduce the filename length

https://github.com/llvm/llvm-project/pull/194662