[PATCH] D153025: AMDGPU: Correctly lower llvm.exp.f32

[Matt Arsenault via Phabricator via llvm-commits]

arsenm created this revision.
arsenm added reviewers: AMDGPU, rampitec, foad, cdevadas, Pierre-vh, jhuber6.
Herald added subscribers: StephenFan, jdoerfert, kerbowa, hiraditya, tpr, dstuttard, yaxunl, jvesely, kzhuravl.
Herald added a project: All.
arsenm requested review of this revision.
Herald added a subscriber: wdng.
Herald added a project: LLVM.

The library expansion has too many paths for all the permutations of
DAZ, unsafe and the 3 exp functions. It's easier to expand it in the
backend when we know all of these things. The library currently misses
the no-infinity check on the overflow, which this handles optimizing
out.

      

Some of the <3 x half> fast tests regress due to vector widening
dropping flags which will be fixed separately.

      

Apparently there is no exp10 intrinsic, but there should be. Adds some
deadish code in preparation for adding one while I'm following along
with the current library expansion.


https://reviews.llvm.org/D153025

Files:
  llvm/docs/AMDGPUUsage.rst
  llvm/docs/LangRef.rst
  llvm/docs/ReleaseNotes.rst
  llvm/include/llvm/CodeGen/GlobalISel/MachineIRBuilder.h
  llvm/lib/Target/AMDGPU/AMDGPUISelLowering.cpp
  llvm/lib/Target/AMDGPU/AMDGPUISelLowering.h
  llvm/lib/Target/AMDGPU/AMDGPULegalizerInfo.cpp
  llvm/lib/Target/AMDGPU/AMDGPULegalizerInfo.h
  llvm/test/CodeGen/AMDGPU/GlobalISel/legalize-fexp.mir
  llvm/test/CodeGen/AMDGPU/llvm.exp.ll

-------------- next part --------------
A non-text attachment was scrubbed...
Name: D153025.531744.patch
Type: text/x-patch
Size: 509832 bytes
Desc: not available
URL: <http://lists.llvm.org/pipermail/llvm-commits/attachments/20230615/62739ae7/attachment-0001.bin>