| commit | 5acacb72256127684847de79d15b6607e869a377 | [log] [tgz] |
|---|---|---|
| author | Quinn Dawkins <quinn.dawkins@gmail.com> | Wed Mar 27 21:09:44 2024 -0400 |
| committer | GitHub <noreply@github.com> | Wed Mar 27 21:09:44 2024 -0400 |
| tree | 315fe6a41c6d0c8291e7aae2311d089b633e7728 | |
| parent | bfdbd166170cc9bdeb6bd645a16f2c8e0929ec9f [diff] |
[Codegen] Fix layout analysis for vector.transpose (#16820) (#16921) Previously transpose analysis would use the same permutation when propagating upwards and downwards through the transpose, leading to layout conflicts and crashes when the transpose was not an involution (i.e. its own inverse). Additionally the permutation logic for the nested layout did not properly invert the permutation when applied to the order. The logic is described in detailed comments by Kunwar, but in short, the sizes are permuted by the permutation, so to map that back to the original layout, we need to apply the inverse permutation to the layout order fields. Co-authored-by: Kunwar Grover <groverkss@gmail.com>
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.