Skip to content
Merged
Show file tree
Hide file tree
Changes from all commits
Commits
File filter

Filter by extension

Filter by extension


Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
48 changes: 0 additions & 48 deletions .github/workflows/release.yaml
Original file line number Diff line number Diff line change
Expand Up @@ -83,51 +83,3 @@ jobs:
path: dist/
- name: publish package distributions to PyPI
uses: pypa/gh-action-pypi-publish@7245906fb61fa33a13554f9d5fb856fe61758086 # v1.14.3

# ------------------------------------

build-oeq-extjax:
name: Build extjax dist
runs-on: ubuntu-latest
steps:
- uses: actions/checkout@3d3c42e5aac5ba805825da76410c181273ba90b1 # v7.0.1
with:
fetch-tags: true
- name: set up Python
uses: actions/setup-python@6d4fd22b20465ccc0f7354476588e8df7c9ca0e0 # v7.1.0
with:
python-version: '3.10'
- name: replace symlinks with copies
run: python3 .github/release.py resolve-symlinks
- name: set version from tag
run: |
python3 -m pip install -r .github/workflows/requirements_release_ci.txt
python3 .github/release.py set-version openequivariance_extjax/pyproject.toml
- name: install dependencies, then build source tarball
run: |
cd openequivariance_extjax
python3 -m pip install build
python3 -m build --sdist
- name: store the distribution packages
uses: actions/upload-artifact@cf430e030ddbb5b0abf93d22962f4752f3646cd9 # v7.0.2
with:
name: oeqjax-package-distributions
path: openequivariance_extjax/dist/

pypi-publish-extjax:
name: Upload extjax to PyPI
runs-on: ubuntu-latest
needs: build-oeq-extjax
environment:
name: pypi
url: https://pypi.org/p/openequivariance_extjax
permissions:
id-token: write
steps:
- name: download the distributions
uses: actions/download-artifact@9000827ccba6bdab643e8b6fd33ac0654aef8333 # v8.0.2
with:
name: oeqjax-package-distributions
path: dist/
- name: publish package distributions to PyPI
uses: pypa/gh-action-pypi-publish@7245906fb61fa33a13554f9d5fb856fe61758086 # v1.14.3
6 changes: 1 addition & 5 deletions .github/workflows/verify_extension_build.yml
Original file line number Diff line number Diff line change
Expand Up @@ -29,16 +29,12 @@ jobs:
sudo apt-get update
sudo apt install nvidia-cuda-toolkit
pip install -r .github/workflows/requirements_cuda_ci.txt
pip install -e "./openequivariance[jax]"
pip install -e "./openequivariance"

- name: Test CUDA extension build via import
run: |
pytest tests/import_test.py

- name: Test JAX extension build
run: |
XLA_DIRECT_DOWNLOAD=1 pip install -e "./openequivariance_extjax" --no-build-isolation

verify_hip_extension:
if: ${{ github.event.label.name == 'ci-ready' || github.event_name != 'pull_request' }}
runs-on: ubuntu-latest
Expand Down
18 changes: 18 additions & 0 deletions CHANGELOG.md
Original file line number Diff line number Diff line change
@@ -1,5 +1,23 @@
## Latest Changes

### v0.8.2 (2026-11-17)

This release eliminates the need
for `openequivariance_extjax` by moving
the JAX extension to a precompiled library
inside a wheeldist. Also introduces
quality-of-life changes to JIT branch.


**Deprecated**:
- JAX users no longer need to
install `openequivariance_extjax` in addition
to the regular `openequivariance` package.

### v0.8.1 (2026-11-10)

This release is a preliminary test of precompiled library wheel distribution.

**Added**:
- Route symmetric contraction's grouped matrix multiplications through PyTorch
on CUDA, HIP, and SYCL, removing direct cuBLAS and rocBLAS dependencies.
Expand Down
16 changes: 1 addition & 15 deletions README.md
Original file line number Diff line number Diff line change
Expand Up @@ -34,21 +34,7 @@ OpenEquivariance now supports Intel XPUs on PyTorch
by compiling SYCL kernels. See our
installation guide for more details.

⭐️ **JAX**: We support JAX for NVIDIA and AMD GPUs.
For NVIDIA GPUs, install it (after installing JAX)
with the following two commands strictly in order:

``` bash
pip install openequivariance[jax]
pip install openequivariance_extjax --no-build-isolation
```

For AMD GPUs:
``` bash
pip install openequivariance[jax]
JAX_HIP=1 pip install openequivariance_extjax --no-build-isolation
```

⭐️ **JAX**: We support JAX for NVIDIA and AMD GPUs.
See the section below for example usage and
our [API page](https://passionlab.github.io/OpenEquivariance/api/) for more details.

Expand Down
1 change: 0 additions & 1 deletion docs/conf.py
Original file line number Diff line number Diff line change
Expand Up @@ -37,7 +37,6 @@
"jax",
"openequivariance._torch.extlib",
"openequivariance.jax.extlib",
"openequivariance_extjax",
"openequivariance.jax.jvp.tp_prim",
"openequivariance.jax.jvp.conv_prim",
"jinja2",
Expand Down
115 changes: 57 additions & 58 deletions docs/installation.rst
Original file line number Diff line number Diff line change
@@ -1,110 +1,109 @@
Installation
==============================

.. admonition:: Deprecation of ``openequivariance_extjax``
:class: tip

Installing ``openequivariance_extjax`` is no longer
required or supported, since the regular
``openequivariance`` package contains the
relevant binaries.


.. toctree::
:maxdepth: 1
:caption: Contents:

You need the following to install OpenEquivariance:

- A Linux system equipped with an NVIDIA / AMD / Intel graphics card.
- Either PyTorch >= 2.10, or
JAX>0.5.0 with CUDA, RocM, or XPU support.
- To build from source: GCC 9+ and the CUDA / HIP toolkit, or the
oneAPI DPC++ compiler (``icpx``) for Intel GPUs. The command
``c++ --version`` should return >= 9.0; see below for details on
setting an alternate compiler.

.. note::

On Intel GPUs the kernels are generated as SYCL and compiled at runtime
through the oneAPI kernel compiler, so ``icpx`` must be on your ``PATH``
(PyTorch locates the SYCL toolchain from it). Tensors passed to the
kernels live on the ``xpu`` device rather than ``cuda``. The precompiled
extension and the JAX frontend both remain CUDA/HIP only.
- Either PyTorch >= 2.10, or JAX >= 0.8.1 with CUDA, ROCm, or XPU support.

.. tab:: PyTorch
.. tab:: CUDA or HIP Wheels

Installation is one easy command, followed by import verification:
Installation is one easy command for all PyTorch, JAX, CUDA, and
HIP platforms:

.. code-block:: bash

pip install openequivariance
python -c "import openequivariance"

OpenEquivariance bundles a precompiled library files for two recent
CUDA versions and a HIP version. If you're building for an architecture
we don't support or need a different CUDA version, you'll need to build
the package from source:
we don't support or need a different CUDA version, see the source
distribution installation instructions.

.. code-block:: bash

pip install --no-binary openequivariance openequivariance

Set the environment variables ``CC`` and ``CXX`` before the pip install
to build with a different compiler:
If you're using JAX, set the environment variable
``OEQ_NOTORCH=1`` to avoid a PyTorch import:

.. code-block:: bash

export CC=/path/to/your/gcc
export CXX=/path/to/your/g++
export OEQ_NOTORCH=1
python -c "import openequivariance.jax"

On Intel GPUs, the extension is still compiled on first import,
which can take a couple of minutes; subsequent imports use the
cached build. Install the extra build dependencies with:
.. tab:: CUDA or HIP Source Distribution

Run

.. code-block:: bash

pip install openequivariance[sycl]
pip install --no-binary openequivariance openequivariance

Optionally, set the environment variables ``CC`` and ``CXX`` before the pip install
to build with a different compiler:

.. tab:: JAX NVIDIA GPUs
.. code-block:: bash

First ensure the appropriate JAX Python
package is installed in your environment. Then
run the following two commands stricly in order:
export CC=/path/to/your/gcc
export CXX=/path/to/your/g++
pip install --no-binary openequivariance openequivariance

.. code-block:: bash

pip install openequivariance[jax]
pip install openequivariance_extjax --no-build-isolation
.. tab:: SYCL

.. tab:: JAX AMD GPUs
You'll need the oneAPI DPC++ compiler (``icpx``) for Intel GPUs.
Kernels are generated as SYCL and compiled at runtime
through the oneAPI kernel compiler, so ``icpx`` must be on your ``PATH``
(PyTorch locates the SYCL toolchain from it). Tensors passed to the
kernels are resident on the ``xpu`` device rather than ``cuda``.

Ensure that JAX is installed correctly with RocM support
before running, in order,
Our SYCL extension is still compiled on first import,
which can take a couple of minutes; subsequent imports use the
cached build. Install via

.. code-block:: bash
.. code-block:: bash

pip install openequivariance[jax]
JAX_HIP=1 pip install openequivariance_extjax --no-build-isolation
pip install "openequivariance[sycl]"
export CC=/path/to/your/cc
export CXX=/path/to/your/c++
python -c "import openequivariance"


.. tab:: Nightly (PT)
.. tab:: Developer Install

.. code-block:: bash
This mode enables fast recompilation for
incremental builds. Build isolation is disabled
in this mode:

pip install "git+https://github.com/PASSIONLab/OpenEquivariance#subdirectory=openequivariance"
.. code-block:: bash

git clone https://github.com/PASSIONLab/OpenEquivariance
cd OpenEquivariance
pip install $(python -c "import tomllib; print(*tomllib.load(open('openequivariance/pyproject.toml', 'rb'))['build-system']['requires'])")
pip install --no-build-isolation -Ceditable.rebuild=true -e ./openequivariance

.. tab:: Nightly (JAX)
.. tab:: Nightly

.. code-block:: bash

pip install "openequivariance[jax] @ git+https://github.com/PASSIONLab/OpenEquivariance#subdirectory=openequivariance"
pip install "git+https://github.com/PASSIONLab/OpenEquivariance#subdirectory=openequivariance_extjax" --no-build-isolation
pip install --force-reinstall "git+https://github.com/PASSIONLab/OpenEquivariance#subdirectory=openequivariance"

# Use the command below for JAX+AMD
# JAX_HIP=1 pip install "git+https://github.com/PASSIONLab/OpenEquivariance#subdirectory=openequivariance_extjax" --no-build-isolation
To install from a specific branch on our repo:

.. code-block:: bash

If you're using JAX, set the environment variable
``OEQ_NOTORCH=1`` to avoid a PyTorch import:

.. code-block:: bash

export OEQ_NOTORCH=1
python -c "import openequivariance.jax"
pip install --force-reinstall "git+https://github.com/PASSIONLab/OpenEquivariance@my-test-branch#subdirectory=openequivariance"


Configurations on Major Platforms
Expand Down
3 changes: 1 addition & 2 deletions docs/tests_and_benchmarks.rst
Original file line number Diff line number Diff line change
Expand Up @@ -35,9 +35,8 @@ To set up an editable install and run our tests, use the following code:
git clone https://github.com/PASSIONLab/OpenEquivariance
cd OpenEquivariance

pip install "./openequivariance[jax]"
pip install "./openequivariance"
pip install "./openequivariance[dev]"
pip install "./openequivariance_extjax" --no-build-isolation

pytest --jax tests/example_test.py
pytest --jax tests/batch_test.py
Expand Down
Loading
Loading