commit | beda464d2ca86d2a6cb3462476566e67326d707c | [log] [tgz] |
---|---|---|
author | MaheshRavishankar <1663364+MaheshRavishankar@users.noreply.github.com> | Thu Jun 30 18:04:13 2022 -0700 |
committer | GitHub <noreply@github.com> | Thu Jun 30 18:04:13 2022 -0700 |
tree | 15a71853e8f1f54aea5855899ff1cfaa79b469cd | |
parent | ca258bcd0dea4d22b4dddc9ef05382d6bd61afd3 [diff] |
Add `flow.dispatch.default_workgroup_count` op. (#9681) This commit adds an operation flow.dispatch.default_workgroup_count that captures the relationship between the workload and workgroup_count at the Flow level. This is populated now at the Flow level and carried through to the HAL layer. There the backends lower this operation to the actual computation that returns the workgroup count. There could be many such ops. The current operation is just reflacting the computation that is done today in the backend. In future more such operations and corresponding lowerings in the backends could be added. With this change, the backend expects the workgroup_count region to be populated always.
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.