| commit | 91ee20f9895c55ddc8fba3f788fc31d854fef3e7 | [log] [tgz] |
|---|---|---|
| author | Lei Zhang <antiagainst@gmail.com> | Tue Feb 13 10:45:48 2024 -0800 |
| committer | GitHub <noreply@github.com> | Tue Feb 13 13:45:48 2024 -0500 |
| tree | 5a495bf0279e365328edeb9bc626e725d13cbb40 | |
| parent | c70bf22b07893542b00afdce800d8c77ec23a19f [diff] |
[rocdl] Create basic code structure for ROCDL CodeGen (#16342) This commit starts the rework of LLVMGPU compiler backend for future production. Specifically it starts a set of new source code files for ROCDL CodeGen configuration. The goal is to start with a minimal working structure and gradually port over relevant pass pipelines with proper target handling and heuristics. So in this commit we only pick the warp reduction pipeline which sort of matches the criteria here. Also we don't shuffle the passes around yet--the new ROCDL library just depends on the existing LLVMGPU library and its passes directly to get started. This is meant to avoid churns and we can do the mechanical change once everything is sorted. Progress towards https://github.com/openxla/iree/issues/16341
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.