| commit | 80e70ca460ba07e57df7da56b9e8302c3e358e9f | [log] [tgz] |
|---|---|---|
| author | Ben Vanik <ben.vanik@gmail.com> | Wed Dec 13 15:34:37 2023 -0800 |
| committer | GitHub <noreply@github.com> | Wed Dec 13 15:34:37 2023 -0800 |
| tree | bde00424f3f42aaca496c2f3188a601c48195ced | |
| parent | 3654fd683c1cbb5df76023dd0880ca164fba55c8 [diff] |
Replacing hal.ex.shared_device with hal.devices.* ops. (#15916) This bumps the HAL version as the existing hal.ex.shared_device op is removed. The runtime HAL module still only reports a single device and the compiler code is still selecting the "default" device (0). Future changes will support multiple `--device=` flags to register multiple available devices in tools and add attributes/lookup for mapping stream/device affinity to device ordinals. The plan is to have a `hal.devices.lookup` op that is hoisted into an initializer to enumerate devices and store the selected ones in globals. The enumeration code will emit a loop over `hal.devices.count` and use `hal.devices.get` and the various device query methods to match the requested devices.
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.
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!
See our website for more information.
IREE is licensed under the terms of the Apache 2.0 License with LLVM Exceptions. See LICENSE for more information.