| commit | 8ee423c8fb4a92112a1731fe082b981479edcb43 | [log] [tgz] |
|---|---|---|
| author | Stella Laurenzo <stellaraccident@gmail.com> | Tue Nov 14 11:29:46 2023 -0800 |
| committer | GitHub <noreply@github.com> | Tue Nov 14 11:29:46 2023 -0800 |
| tree | 6c33a03b5e3caa8cc6b5f2a991b1efaefd7c8119 | |
| parent | 35f1884c9eb01e50b033e70a23cc85e9d59a97b4 [diff] |
Add optional CMake presets to the repo. (#15581) Since people have complained about this in the past with respect to IDEs doing weird things, I created these in a sub-directory with the intention to just have a small include file in the root. Leaving that up to the user to do until we can test it more with real dev flows. With this, I was able to get: * Minimal dev down to 3800 ninja actions to build iree-compile. * ~5900 ninja actions for Turbine with a host-only CPU backend. There are on the order of several hundred C++ files that I expect will be eliminated once more of the compiler components are turned into plugins with their dependent pieces contained (i.e. as opposed to having dependencies for all targets implicitly included in top-level registrations). Currently, this includes common code needed for SPIRV, CPU compilation, various GPUs, etc (even when the corresponding target is not enabled). Further, I will do more surgery to make the clang build optional for people who can satisfy the dep with a host clang. This will drop true min-spec CPU builds by ~1000 targets. Progress on #5804 skip-ci: none
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.