Google Git
Sign in
opensecura / 3p / openxla / iree / b5af9963ff9ba8dcf872434a3a7b8c9bb999db0e
commitb5af9963ff9ba8dcf872434a3a7b8c9bb999db0e[log] [tgz]
authorBenoit Jacob <jacob.benoit.1@gmail.com>Fri Mar 08 11:41:27 2024 -0500
committerGitHub <noreply@github.com>Fri Mar 08 11:41:27 2024 -0500
tree6ae17a93e29c71940a490d0bf3ef5d4bb337d3ab
parent47dfaa54c2cae59323a43d7c001a693090cd1c78 [diff]
Prefer broadcasting RHS over LHS in AVX-512 multiply-accumulate instructions (#16709)

Some AVX-512 multiply-accumulate instructions have variants where the
RHS is a `m32bcst`, which is, a 32-bit memory operand with implicit
broadcast semantics. These save having to use a separate broadcast
instruction so they are preferable.

Depending on each compiler version and each intrinsic, it may or may not
make a difference which operand is passed as LHS/RHS, that is, some
compilers are able to swap LHS<->RHS as needed to make this happen.

Compiler Explorer: https://godbolt.org/z/eh81x5qT8

Summary:

Compiler | `vfmadd231ps_broadcast_rhs` | `vfmadd231ps_broadcast_lhs` |
`vpdpwssd_broadcast_rhs` | `vpdpwssd_broadcast_lhs`
--- | --- | --- | --- | ---
Clang trunk | ✅ | ✅ | ✅ | ✅
Clang 17 | ✅ | ✅ | ✅ | ❌
GCC trunk | ✅ | ✅ | ❌ | ❌
MSVC latest | ❌ | ❌ | ❌ | ❌

So... this PR helps at least on some Clang versions (which happen to
include my native Clang-16 toolchain here on Ubuntu).
2 files changed
tree: 6ae17a93e29c71940a490d0bf3ef5d4bb337d3ab
  1. .devcontainer/
  2. .github/
  3. build_tools/
  4. compiler/
  5. docs/
  6. experimental/
  7. integrations/
  8. lib/
  9. llvm-external-projects/
  10. runtime/
  11. samples/
  12. tests/
  13. third_party/
  14. tools/
  15. .bazel_to_cmake.cfg.py
  16. .bazelignore
  17. .bazelrc
  18. .bazelversion
  19. .clang-format
  20. .dockerignore
  21. .git-blame-ignore-revs
  22. .gitignore
  23. .gitmodules
  24. .yamllint.yml
  25. AUTHORS
  26. BUILD.bazel
  27. CITATION.cff
  28. CMakeLists.txt
  29. configure_bazel.py
  30. CONTRIBUTING.md
  31. LICENSE
  32. README.md
  33. WORKSPACE
README.md

IREE: Intermediate Representation Execution Environment

IREE (Intermediate Representation Execution Environment, pronounced as “eerie”) is an MLIR-based end-to-end compiler and runtime that lowers Machine Learning (ML) models to a unified IR that scales up to meet the needs of the datacenter and down to satisfy the constraints and special considerations of mobile and edge deployments.

See our website for project details, user guides, and instructions on building from source.

CI Status IREE Discord Status

Project Status

IREE is still in its early phase. We have settled down on the overarching infrastructure and are actively improving various software components as well as project logistics. It is still quite far from ready for everyday use and is made available without any support at the moment. With that said, we welcome any kind of feedback on any communication channels!

Communication Channels

Related Project Channels

  • MLIR topic within LLVM Discourse: IREE is enabled by and heavily relies on MLIR. IREE sometimes is referred to in certain MLIR discussions. Useful if you are also interested in MLIR evolution.

Architecture Overview

IREE Architecture IREE Architecture

See our website for more information.

Presentations and Talks

License

IREE is licensed under the terms of the Apache 2.0 License with LLVM Exceptions. See LICENSE for more information.