| commit | eeb6e80d4ade331a4271f92ea776a7afcbdc9ecb | [log] [tgz] |
|---|---|---|
| author | MaheshRavishankar <1663364+MaheshRavishankar@users.noreply.github.com> | Wed Aug 09 11:47:54 2023 -0700 |
| committer | GitHub <noreply@github.com> | Wed Aug 09 11:47:54 2023 -0700 |
| tree | e58f68bf9b5ef86ba3c0308d81032c324389728f | |
| parent | dd89a32697a12143e79ecd80b18c52bcc95a5238 [diff] |
Add a preprocessing pass to move entire function into a single dispatch. (#14578) For cases where the model is very small and does not have much concurrency, it is better to move the entire function body into a single dispatch. Eventually the default heuristics can probably figure out when a model is "too small", but for now this PR adds a pass to move the entire function body into a single dispatch to use as a way to find codegen issues such an approach throws up, and also to experiment with different heuristics needed to find such dispatches automatically.
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.