forked from OSchip/llvm-project
503deec218
As disscussed in post-commit review starting with
https://reviews.llvm.org/D84108#2227365
while this appears to be mostly a win overall, especially code-size-wise,
this appears to shake //certain// code pattens in a way that is extremely
unfavorable for performance (+30% runtime regression)
on certain CPU's (i personally can't reproduce).
So until the behaviour is better understood, and a path forward is mapped,
let's back this out for now.
This reverts commit
|
||
|---|---|---|
| .. | ||
| X86 | ||
| 2010-03-22-empty-baseclass.ll | ||
| PR6627.ll | ||
| basic.ll | ||
| bitfield-bittests.ll | ||
| d83507-knowledge-retention-bug.ll | ||
| gdce.ll | ||
| globalaa-retained.ll | ||
| inlining-alignment-assumptions.ll | ||
| lifetime-sanitizer.ll | ||
| loop-rotation-vs-common-code-hoisting.ll | ||
| min-max-abs-cse.ll | ||
| minmax.ll | ||
| reassociate-after-unroll.ll | ||
| rotate.ll | ||
| scev-custom-dl.ll | ||
| scev.ll | ||
| simplifycfg-options.ll | ||
| two-shifts-by-sext.ll | ||
| unsigned-multiply-overflow-check.ll | ||
| vector-trunc.ll | ||