mpi4jax

mpi4jax

|JOSS paper| |PyPI Version| |Conda Version| |Tests| |codecov| |Documentation Status|

mpi4jax enables zero-copy, multi-host communication of JAX <https://jax.readthedocs.io/>_ arrays, even from jitted code and from GPU memory.

But why?

The JAX framework has great performance for scientific computing workloads <https://github.com/dionhaefner/pyhpc-benchmarks>, but its multi-host capabilities <https://jax.readthedocs.io/en/latest/jax.html#jax.pmap> are still limited.

With mpi4jax, you can scale your JAX-based simulations to entire CPU and GPU clusters (without ever leaving jax.jit).

In the spirit of differentiable programming, mpi4jax also supports differentiating through some MPI operations.

Installation

mpi4jax is available through pip and conda:

.. code:: bash

$ pip install mpi4jax # Pip $ conda install -c conda-forge mpi4jax # conda

Depending on the different jax backends you want to use, you can install mpi4jax in the following way

.. code:: bash

pip install 'jax[cpu]'

$ pip install mpi4jax

pip install -U 'jax[cuda12]'

$ pip install cython $ pip install mpi4jax --no-build-isolation

pip install -U 'jax[cuda12_local]'

$ CUDA_ROOT=XXX pip install mpi4jax

(for more informations on jax GPU distributions, see the JAX installation instructions <https://github.com/google/jax#installation>_)

In case your MPI installation is not detected correctly, it can help to install mpi4py separately <https://mpi4py.readthedocs.io/en/stable/install.html>_. When using a pre-installed mpi4py, you must use --no-build-isolation when installing mpi4jax:

.. code:: bash

if mpi4py is already installed

$ pip install cython $ pip install mpi4jax --no-build-isolation

Our documentation includes some more advanced installation examples. <https://mpi4jax.readthedocs.io/en/latest/installation.html>_

Example usage

.. code:: python

from mpi4py import MPI import jax import jax.numpy as jnp import mpi4jax

comm = MPI.COMM_WORLD rank = comm.Get_rank()

@jax.jit def foo(arr): arr = arr + rank arr_sum, _ = mpi4jax.allreduce(arr, op=MPI.SUM, comm=comm) return arr_sum

a = jnp.zeros((3, 3)) result = foo(a)

if rank == 0: print(result)

Running this script on 4 processes gives:

.. code:: bash

$ mpirun -n 4 python example.py [[6. 6. 6.] [6. 6. 6.] [6. 6. 6.]]

allreduce is just one example of the MPI primitives you can use. See all supported operations here <https://mpi4jax.readthedocs.org/en/latest/api.html>_.

Community guidelines

If you have a question or feature request, or want to report a bug, feel free to open an issue <https://github.com/mpi4jax/mpi4jax/issues>_.

We welcome contributions of any kind through pull requests <https://github.com/mpi4jax/mpi4jax/pulls>. For information on running our tests, debugging, and contribution guidelines please refer to the corresponding documentation page <https://mpi4jax.readthedocs.org/en/latest/developers.html>.

How to cite

If you use mpi4jax in your work, please consider citing the following article:

::

@article{mpi4jax, doi = {10.21105/joss.03419}, url = {https://doi.org/10.21105/joss.03419}, year = {2021}, publisher = {The Open Journal}, volume = {6}, number = {65}, pages = {3419}, author = {Dion Häfner and Filippo Vicentini}, title = {mpi4jax: Zero-copy MPI communication of JAX arrays}, journal = {Journal of Open Source Software} }

.. |Tests| image:: https://github.com/mpi4jax/mpi4jax/workflows/Tests/badge.svg :target: https://github.com/mpi4jax/mpi4jax/actions?query=branch%3Amaster .. |codecov| image:: https://codecov.io/gh/mpi4jax/mpi4jax/branch/master/graph/badge.svg :target: https://codecov.io/gh/mpi4jax/mpi4jax .. |PyPI Version| image:: https://img.shields.io/pypi/v/mpi4jax :target: https://pypi.org/project/mpi4jax/ .. |Conda Version| image:: https://img.shields.io/conda/vn/conda-forge/mpi4jax.svg :target: https://anaconda.org/conda-forge/mpi4jax .. |Documentation Status| image:: https://readthedocs.org/projects/mpi4jax/badge/?version=latest :target: https://mpi4jax.readthedocs.io/en/latest/?badge=latest .. |JOSS paper| image:: https://joss.theoj.org/papers/10.21105/joss.03419/status.svg :target: https://doi.org/10.21105/joss.03419