xDSL: A Python-native SSA Compiler Framework
xDSL is a Python-native framework for building compiler infrastructure. It provides SSA-based intermediate representations (IRs) and Pythonic APIs to define, assemble, and optimize custom IRs—all with seamless compatibility with MLIR from the LLVM project.
Inspired by MLIR, xDSL enables smooth translation of programs and abstractions between frameworks. This lets users prototype compilers entirely in Python, while still accessing MLIR's powerful optimization and code generation pipeline. All IRs in xDSL employ a unified SSA-based data structure, with regions and basic blocks, making it easy to write generic analyses and transformation passes.
xDSL supports assembling compilers from predefined or custom IRs, and organizing transformations across a multi-level IR stack. This layered approach enables abstraction-specific optimization passes, similar to the architecture of projects like Devito, PSyclone, and Firedrake.
In short, xDSL makes it possible to:
- Prototype compilers quickly in Python
- Build DSLs with custom IRs
- Run analyses and transformations with simple scripts
- Interoperate smoothly with MLIR and benefit from LLVM's backend
Contents
Installation
To contribute to xDSL, follow the xDSL Developer Setup Guide.
To use xDSL as part of a larger project for developing your own compiler, just install xDSL via pip:
pip install xdsl
To quickly install xDSL for development and contribution purposes, use:
pip install xdsl[dev]
This may be useful for projects wanting to replicate the xDSL testing setup.
Note: This version of xDSL is validated against a specific MLIR version, interoperability with other versions is not guaranteed. The supported MLIR version is 22.1.2.
[!IMPORTANT]
Experimental Pyright Features
xDSL currently relies on an experimental feature of Pyright called TypeForm. TypeForm is in discussion and will likely land in some future version of Python.
For xDSL to type check correctly using Pyright, please add this to your
pyproject.toml:[tool.pyright] enableExperimentalFeatures = true
Subprojects With Extra Dependencies
xDSL has a number of subprojects, some of which require extra dependencies. To keep the set of dependencies to a minimum, these extra dependencies have to be specified explicitly, e.g. by using:
pip install xdsl[gui] # or [jax], [riscv]
Getting Started
Check out the dedicated Getting Started guide for a comprehensive tutorial.
To get familiar with xDSL, we recommend starting with our tutorial notebooks. The notebooks provide hands-on examples and documentation of xDSL's core concepts: data structures, the Python-embedded abstraction definition language, and end-to-end custom compilers construction, like a database compiler. There also exists a small documentation showing how to connect xDSL with MLIR for users interested in that use case.
We provide a Makefile containing a lot of common tasks, which might provide an overview of common actions.
Discussion
You can also join the discussion at our Zulip chat room, kindly supported by community hosting from Zulip.
7 maintainers
Download files
Download the file for your platform. If you're not sure which to choose, learn more about installing packages.
Source Distribution
Built Distribution
Filter files by name, interpreter, ABI, and platform.
If you're not sure about the file name format, learn more about wheel file names.
Copy a direct link to the current filters
File details
Details for the file xdsl-0.70.0.tar.gz.
File metadata
- Download URL: xdsl-0.70.0.tar.gz
- Upload date:
- Size: 4.3 MB
- Tags: Source
- Uploaded using Trusted Publishing? No
- Uploaded via:
twine/7.0.0 CPython/3.11.16
File hashes
| Algorithm | Hash digest | |
|---|---|---|
| SHA256 |
86295a650cca786f2da2b5ec23953c777e6cd2b90b1315a39bd4a123742666a9
|
|
| MD5 |
3b36609afbabfe0cb5a91a6dbfdc4b11
|
|
| BLAKE2b-256 |
ce67a303961d494bb3f95f5d030632912fde66bb53e5743ef7b9a322b0f12e7b
|
File details
Details for the file xdsl-0.70.0-py3-none-any.whl.
File metadata
- Download URL: xdsl-0.70.0-py3-none-any.whl
- Upload date:
- Size: 4.5 MB
- Tags: Python 3
- Uploaded using Trusted Publishing? No
- Uploaded via:
twine/7.0.0 CPython/3.11.16
File hashes
| Algorithm | Hash digest | |
|---|---|---|
| SHA256 |
f80a5356470af0dac46083ddde5bf1896271f19676aac7fb04ba44590d47c52b
|
|
| MD5 |
2c2685febbbcc8b50390154ee59c6483
|
|
| BLAKE2b-256 |
d489049f192e68cc95a69334ffad9f5af93f23b7b46a104dd45fa733b5ba5ef7
|
