| commit | 7acfdb79b2628e58341700ca47e561262241b5d5 | [log] [tgz] |
|---|---|---|
| author | Ben Vanik <ben.vanik@gmail.com> | Thu Aug 17 16:00:26 2023 -0700 |
| committer | GitHub <noreply@github.com> | Thu Aug 17 23:00:26 2023 +0000 |
| tree | abbfd00229ac3367a3b29fb2e523248b17e5d714 | |
| parent | ce0a6127d433eebaadbde8030c0b1fd685479c3e [diff] |
Adding MTLBuffer import from host and device pointers. (#14703) Unfortunately Metal has some strict requirements around what it can import even on unified memory systems and our tooling can't trigger this path yet (results of malloc can never be imported). There may be some ways around this with memfd-like insanity but for now being able to zero-copy pass in MTLBuffers allocated externally is the best bet. Progress on #14607. Fixes #14608.
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.