[all-commits] [llvm/llvm-project] 711419: [flang] Enable loop-versioning for slices. (#120344)

[Slava Zakharin via All-commits]

  Branch: refs/heads/main
  Home:   https://github.com/llvm/llvm-project
  Commit: 711419e3025678511e3d26c4c30d757f9029d598
      https://github.com/llvm/llvm-project/commit/711419e3025678511e3d26c4c30d757f9029d598
  Author: Slava Zakharin <szakharin at nvidia.com>
  Date:   2024-12-23 (Mon, 23 Dec 2024)

  Changed paths:
    M flang/lib/Optimizer/Transforms/LoopVersioning.cpp
    M flang/test/Transforms/loop-versioning.fir

  Log Message:
  -----------
  [flang] Enable loop-versioning for slices. (#120344)

Loops resulting from array expressions like array(:,i)
may be versioned for the unit stride of the innermost dimension,
when the initial array is an assumed-shape array (which are contiguous
in many Fortran programs).
This speeds up facerec for about 12% due to further vectorization
of the innermost loop produced for the total SUM reduction.



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