David Majnemer 8b680c27be [SLPVectorizer] Vectorizing the libm sqrt to llvm's sqrt intrinsic requires nnan ...

To quote the langref "Unlike sqrt in libm, however, llvm.sqrt has
undefined behavior for negative numbers other than -0.0 (which allows
for better optimization, because there is no need to worry about errno
being set). llvm.sqrt(-0.0) is defined to return -0.0 like IEEE sqrt."

This means that it's unsafe to replace sqrt with llvm.sqrt unless the
call is annotated with nnan.

Thanks to Hal Finkel for pointing this out!

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@265521 91177308-0d34-0410-b5e6-96231b3b80d8

2016-04-06 07:04:53 +00:00

..

AArch64

Reapply commit r259357 with a fix for PR26629

2016-02-18 14:14:40 +00:00

AMDGPU

[SLPVectorizer] Try different vectorization factors for store chains

2015-07-08 23:40:55 +00:00

ARM

[opaque pointer type] Add textual IR support for explicit type parameter to load instruction

2015-02-27 21:17:42 +00:00

X86

[SLPVectorizer] Vectorizing the libm sqrt to llvm's sqrt intrinsic requires nnan

2016-04-06 07:04:53 +00:00

XCore

[opaque pointer type] Add textual IR support for explicit type parameter to load instruction

2015-02-27 21:17:42 +00:00