[all-commits] [llvm/llvm-project] 963ff1: [mlir] lower min/maxnum to libdevice calls (#127323)

[ftynse via All-commits]

  Branch: refs/heads/main
  Home:   https://github.com/llvm/llvm-project
  Commit: 963ff1c3051a8c413a04fff27ba60cf93e48f310
      https://github.com/llvm/llvm-project/commit/963ff1c3051a8c413a04fff27ba60cf93e48f310
  Author: Oleksandr "Alex" Zinenko <git at ozinenko.com>
  Date:   2025-02-15 (Sat, 15 Feb 2025)

  Changed paths:
    M mlir/lib/Conversion/GPUToNVVM/LowerGpuOpsToNVVMOps.cpp
    A mlir/test/Conversion/GPUToNVVM/gpu-arith-ops-to-nvvm.mlir

  Log Message:
  -----------
  [mlir] lower min/maxnum to libdevice calls (#127323)

Introduce lowering from arith.minnum/maxxnum operations to the
corresponding Nvidia libdevice calls. This requires to reorder pattern
population methods so that the libdevice-targeting patterns are
prioritized over default patterns targeting LLVM IR intrinsics from the
Arith dialect. The tests are placed into a separate file as the existing
gpu-to-nvvm.mlir files has a mode that forces Arith dialect operations
to be preserved as is without using a separate FileCheck tag to
differentiate.

Co-authored-by: William Moses <gh at wsmoses.com>



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