[all-commits] [llvm/llvm-project] 574701: DAG: Lower is.fpclass fcSubnormal|fcZero to fabs(x...

[Matt Arsenault via All-commits]

  Branch: refs/heads/users/arsenm/dag-lower-isfpclass-normal-or-zero-to-fcmp-fabs-smallest-normal
  Home:   https://github.com/llvm/llvm-project
  Commit: 574701f8aaad9305922c39e278e1dfd1acc26b75
      https://github.com/llvm/llvm-project/commit/574701f8aaad9305922c39e278e1dfd1acc26b75
  Author: Matt Arsenault <Matthew.Arsenault at amd.com>
  Date:   2024-07-24 (Wed, 24 Jul 2024)

  Changed paths:
    M llvm/lib/CodeGen/SelectionDAG/TargetLowering.cpp
    M llvm/test/CodeGen/X86/is_fpclass.ll

  Log Message:
  -----------
  DAG: Lower is.fpclass fcSubnormal|fcZero to fabs(x) < smallest_normal

Produces better code on x86_64 only in the unordered case. Not
sure what the exact condition should be to avoid the regression. Free
fabs might do it, or maybe requires legality checks for the alternative
integer expansion.



To unsubscribe from these emails, change your notification settings at https://github.com/llvm/llvm-project/settings/notifications