| commit | 096ad43f8e50bf1de93e3c575113b78ca590fe25 | [log] [tgz] |
|---|---|---|
| author | Nicolas Vasilache <nicolasvasilache@users.noreply.github.com> | Mon Nov 14 05:03:25 2022 -0800 |
| committer | GitHub <noreply@github.com> | Mon Nov 14 05:03:25 2022 -0800 |
| tree | 46ffad848c6dd7ff85f0ce4680dc090ba1b455d1 | |
| parent | 866ca5aa0ec8f8b15972d6c3ac5279ec12e0f887 [diff] |
Introduce a transform.do_not_dce_operands (#11147) The new op is introduced in LinalgExt for now as it cannot be dynamically registered into the transform dialect atm. With this op, canonicalized_sequence does not DCE the SSA values introduced by ConfigExtractPartOp. This is necessary to allow the next part of the config to be extracted without either DCE or dropped attributes under transforms limiting us. With the existing ListenerGreedyPatternRewriteDriver.cpp, the new op would still be DCE'ed (!!!). So we also adapt the code to its recent upstream evolution.
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.