Allow flags to be set with greater flexibility (#17659)

Changes to the python binding to allow iree.runtime.flags.parse_flags to
take effect at times other than before the first time a driver is
created. Also includes fixes for bugs exposed during the development of
this feature.

- Added "internal" API functions `create_hal_driver()` and
`clear_hal_driver_cache()` to create a driver object independent of the
cache, and to clear the cache, respectively
- Added `HalDriver` class implementation functions for the above new API
functions. Refactored class to share as much common code as possible.
- Factored out driver URI processing into its own nested class for
easier handling of URI components
- Fixed dangling pointer bug. In the C layer flags are being kept by
reference as string views, requiring the caller to keep the original
flag strings (argc, argv) around for as long as the flags are being
used. However, the python binding was using a local variable for those
strings, letting them go out of scope and causing garbage values later
on. The fix is to move the strings to a file scope variable. Flag
handling does not appear to be getting used in a multi-threaded
environment, as other aspects of flag handling use static variables with
no mutex guarding that I could find.
- Fixed runtime assert in Windows debug build for the improper use of
std::vector<>::front() on an empty vector. The code never used the value
of front(), as it was guarded by a check for the vector's size, but the
assert prevents the debug build from running.

---------

Signed-off-by: Dave Liddell <dave.liddell@amd.com>
Signed-off-by: daveliddell <dave.liddell@amd.com>
5 files changed
tree: 24629e88c2699b82654f8c7edfa4ce6602091a8a
  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. .gitattributes
  23. .gitignore
  24. .gitmodules
  25. .pre-commit-config.yaml
  26. .yamllint.yml
  27. AUTHORS
  28. BUILD.bazel
  29. CITATION.cff
  30. CMakeLists.txt
  31. configure_bazel.py
  32. CONTRIBUTING.md
  33. LICENSE
  34. MAINTAINERS.md
  35. README.md
  36. RELEASING.md
  37. 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

Community meeting recordings: IREE YouTube channel

  • 2021-06-09: IREE Runtime Design Tech Talk (recording and slides)
  • 2020-08-20: IREE CodeGen: MLIR Open Design Meeting Presentation (recording and slides)
  • 2020-03-18: Interactive HAL IR Walkthrough (recording)
  • 2020-01-31: End-to-end MLIR Workflow in IREE: MLIR Open Design Meeting Presentation (recording and slides)

License

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