[all-commits] [llvm/llvm-project] 2c0e09: clang/HIP: Do not call ocml in scalbln implementat...

[Matt Arsenault via All-commits]

  Branch: refs/heads/users/arsenm/clang/hip-fix-scalblnf-implementation
  Home:   https://github.com/llvm/llvm-project
  Commit: 2c0e09d75f4bdad1fbcf052c634ab5a16a54acac
      https://github.com/llvm/llvm-project/commit/2c0e09d75f4bdad1fbcf052c634ab5a16a54acac
  Author: Matt Arsenault <Matthew.Arsenault at amd.com>
  Date:   2025-03-04 (Tue, 04 Mar 2025)

  Changed paths:
    M clang/lib/Headers/__clang_hip_math.h
    M clang/lib/Headers/__clang_hip_runtime_wrapper.h
    M clang/test/Headers/__clang_hip_math.hip

  Log Message:
  -----------
  clang/HIP: Do not call ocml in scalbln implementations

I do not understand why this was calling the float version with
an implicit cast from the long. Just clamp to the bounds of int,
and use the generic ldexp (this is also how musl does it (except
scalbnf is the base implementation there).

Somehow INT_MIN was also not defined, so deal with that.



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