commit | 66162780a42fc179320522c7ce970efbbb5f0843 | [log] [tgz] |
---|---|---|
author | Ben Vanik <ben.vanik@gmail.com> | Mon May 16 13:03:19 2022 -0700 |
committer | GitHub <noreply@github.com> | Mon May 16 13:03:19 2022 -0700 |
tree | 25e33bece126641c16f8fee7b938d9794e686824 | |
parent | 4614568980468f3241780e9c4df2b3e68f457064 [diff] |
Adding simple bytecode versioning. (#9133) This allows for minor version bumps for backward-compatible changes (allowing an older vmfb to be loaded on a newer runtime) and major version bumps (disallowing older vmfbs from being loaded on a newer runtime). This first version 0.0 allows all existing modules to be loaded. Future bumps will result in errors like: ``` E D:\Dev\iree\runtime\src\iree\vm\bytecode_module.c:209: INVALID_ARGUMENT; bytecode version mismatch; runtime supports 0.0, module has 0.1 ``` Note that this only versions the bytecode - the module itself is already versioned with flatbuffers and doesn't need any version numbers. This will also allow us to embed bytecode in other containers (ELF/etc) and reuse the same versioning information.
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.