Matt Arsenault 30ba76520a Fix asserts in AMDGCN fmed3 folding by handling more cases of NaN ...

Better NaN handling for AMDGCN fmed3.

All operands are checked for NaN now. The checks
were moved before the canonicalization to provide
a better mapping from fclamp. Changed the behaviour
of fmed3(x,y,NaN) to return max(x,y) instead of
min(x,y) in light of this. Updated tests as a result
and added some new cases to cover the fix.

Patch by Alan Baker

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@336375 91177308-0d34-0410-b5e6-96231b3b80d8

2018-07-05 17:05:36 +00:00

..

AggressiveInstCombine

Move Analysis/Utils/Local.h back to Transforms

2018-06-04 21:23:21 +00:00

Coroutines

Move Analysis/Utils/Local.h back to Transforms

2018-06-04 21:23:21 +00:00

Hello

Add auto-exporting of symbols from tools so that plugins work on Windows

2016-05-26 11:16:43 +00:00

InstCombine

Fix asserts in AMDGCN fmed3 folding by handling more cases of NaN

2018-07-05 17:05:36 +00:00

Instrumentation

[HWASan] Do not retag allocas before return from the function.

2018-06-29 20:20:17 +00:00

IPO

[UnrollAndJam] New Unroll and Jam pass

2018-07-01 12:47:30 +00:00

ObjCARC

ProvenanceAnalysis: Store WeakTrackingVH instead of Value* in UnderlyingValue Cache.

2018-06-21 05:14:00 +00:00

Scalar

[PM/LoopUnswitch] Fix PR37651 by correctly invalidating SCEV when

2018-07-03 09:13:27 +00:00

Utils

[DebugInfo] Corrections for salvageDebugInfo

2018-07-03 11:29:00 +00:00

Vectorize

[SLPVectorizer] Begin abstracting InstructionsState alternate matching away from opcodes. NFCI.

2018-07-05 12:30:44 +00:00

CMakeLists.txt

Another try to commit 323321 (aggressive instruction combine).

2018-01-25 12:06:32 +00:00

LLVMBuild.txt

Another try to commit 323321 (aggressive instruction combine).

2018-01-25 12:06:32 +00:00