[all-commits] [llvm/llvm-project] 80d327: [Hexagon] Enable soft bf16 in hexagon (#167924)

[Fateme Hosseini via All-commits]

  Branch: refs/heads/main
  Home:   https://github.com/llvm/llvm-project
  Commit: 80d327e84606a99fd2c0b3f6ddb37f32ed89b997
      https://github.com/llvm/llvm-project/commit/80d327e84606a99fd2c0b3f6ddb37f32ed89b997
  Author: Fateme Hosseini <quic_fhossein at quicinc.com>
  Date:   2025-11-19 (Wed, 19 Nov 2025)

  Changed paths:
    M clang/lib/Basic/Targets/Hexagon.cpp
    M clang/lib/Basic/Targets/Hexagon.h
    M llvm/lib/Target/Hexagon/HexagonCallingConv.td
    M llvm/lib/Target/Hexagon/HexagonISelLowering.cpp
    M llvm/lib/Target/Hexagon/HexagonISelLoweringHVX.cpp
    M llvm/lib/Target/Hexagon/HexagonPatterns.td
    M llvm/lib/Target/Hexagon/HexagonPatternsHVX.td
    M llvm/lib/Target/Hexagon/HexagonRegisterInfo.td
    M llvm/lib/Target/Hexagon/HexagonSubtarget.h
    A llvm/test/CodeGen/Hexagon/bfloat.ll
    A llvm/test/CodeGen/Hexagon/bfloat_vec.ll

  Log Message:
  -----------
  [Hexagon] Enable soft bf16 in hexagon (#167924)

This patch adds:
1. Support to recognize bf16 type in the frontend and isel/abi support
for scalar bf16 programs
Limitations: fp_to_bf16 is being generated with a tablegen pattern
instead of lowering via expansion. This is because we do not have
support for fcanonincalize instruction which should prevent an SNaN
being converted to an infinity due to truncation.

2. Vector codegen support for bf16

Patch By: Fateme Hosseini

Co-authored-by: Muntasir Mallick <quic_mallick at quicinc.com>
Co-authored-by: Muntasir Mallick <mallick at qti.qualcomm.com>
Co-authored-by: Kaushik Kulkarni <quic_kauskulk at quicinc.com>



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