| commit | 84b4802c5e505c1a1991aa4893a4bb6be6fae043 | [log] [tgz] |
|---|---|---|
| author | Nicolas Vasilache <nicolasvasilache@users.noreply.github.com> | Mon Mar 06 10:07:59 2023 +0100 |
| committer | GitHub <noreply@github.com> | Mon Mar 06 09:07:59 2023 +0000 |
| tree | 8d290d0f96b7e31c550402abcd581fa6507883f1 | |
| parent | 769ffda7c2346e3bba26918f7bbd132bcde5101d [diff] |
Fully retire CanonicalizedSequenceOp (#12467)
This revision fully retires the CanonicalizedSequenceOp and makes the
need for canonicalization / tiling_canonicalization / licm / singleIterationLoopPromotion
explicit in the transform IR.
This removes automatic blanket pattern applications which is expected to be beneficial for:
1. reducing compile time
2. reducing surprise (explicit being better than implicit)
3. help surface issues similar to #12444
In the process, plumb the isolatedFromAbove through the API to make the
requirement explicit until we are able to remove it in some future (i.e. with better targeted
canonicalization and greedy pattern rewriter that does not require an isolated from above).
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.