[PATCH] D87391: [Intrinsics] define semantics for experimental fmax/fmin vector reductions

[Sanjay Patel via Phabricator via llvm-commits]

spatel updated this revision to Diff 290773.
spatel added a comment.

Patch updated:
Change the neutral element for min/max reductions in WidenVecOp_VECREDUCE() from INF to quiet NaN. 
We're probably missing generic and/or target-specific simplifications for those patterns, but I think that can be an enhancement.


CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D87391/new/

https://reviews.llvm.org/D87391

Files:
  llvm/docs/LangRef.rst
  llvm/include/llvm/CodeGen/BasicTTIImpl.h
  llvm/lib/CodeGen/ExpandReductions.cpp
  llvm/lib/CodeGen/SelectionDAG/LegalizeVectorTypes.cpp
  llvm/lib/CodeGen/SelectionDAG/TargetLowering.cpp
  llvm/lib/Target/AArch64/AArch64ISelLowering.cpp
  llvm/test/CodeGen/AArch64/vecreduce-fmax-legalization-nan.ll
  llvm/test/CodeGen/AArch64/vecreduce-fmax-legalization.ll
  llvm/test/CodeGen/Generic/expand-experimental-reductions.ll
  llvm/test/CodeGen/Thumb2/mve-vecreduce-fminmax.ll
  llvm/test/CodeGen/Thumb2/mve-vecreduce-loops.ll
  llvm/test/CodeGen/X86/vector-reduce-fmax-nnan.ll
  llvm/test/CodeGen/X86/vector-reduce-fmax.ll
  llvm/test/CodeGen/X86/vector-reduce-fmin-nnan.ll
  llvm/test/CodeGen/X86/vector-reduce-fmin.ll

-------------- next part --------------
A non-text attachment was scrubbed...
Name: D87391.290773.patch
Type: text/x-patch
Size: 218602 bytes
Desc: not available
URL: <http://lists.llvm.org/pipermail/llvm-commits/attachments/20200909/97faff94/attachment-0001.bin>