Add a link to Intel plugin for JAX

This commit is contained in:
minigoel 2024-10-28 10:47:59 -07:00
parent 6f3c01238e
commit 68428488c8
2 changed files with 14 additions and 0 deletions

View File

@ -390,6 +390,7 @@ Some standouts:
| Google TPU | yes | n/a | n/a | n/a | n/a | n/a |
| AMD GPU | yes | no | experimental | n/a | no | no |
| Apple GPU | n/a | no | n/a | experimental | n/a | n/a |
| Intel GPU | experimental | n/a | n/a | n/a | no | no |
### Instructions
@ -401,6 +402,7 @@ Some standouts:
| Google TPU | `pip install -U "jax[tpu]" -f https://storage.googleapis.com/jax-releases/libtpu_releases.html` |
| AMD GPU (Linux) | Use [Docker](https://hub.docker.com/r/rocm/jax-community/tags), [pre-built wheels](https://github.com/ROCm/jax/releases), or [build from source](https://jax.readthedocs.io/en/latest/developer.html#additional-notes-for-building-a-rocm-jaxlib-for-amd-gpus). |
| Mac GPU | Follow [Apple's instructions](https://developer.apple.com/metal/jax/). |
| Intel GPU | Follow [Intel's instructions](https://github.com/intel/intel-extension-for-openxla/blob/main/docs/acc_jax.md). |
See [the documentation](https://jax.readthedocs.io/en/latest/installation.html)
for information on alternative installation strategies. These include compiling

View File

@ -35,6 +35,7 @@ The table below shows all supported platforms and installation options. Check if
| Google Cloud TPU | {ref}`yes <install-google-tpu>` | n/a | n/a | n/a | n/a | n/a |
| AMD GPU | {ref}`experimental <install-amd-gpu>` | no | {ref}`experimental <install-mac-gpu>` | n/a | no | no |
| Apple GPU | n/a | no | n/a | {ref}`experimental <install-mac-gpu>` | n/a | n/a |
| Intel GPU | {ref}`experimental <install-intel-gpu>`| n/a | n/a | n/a | no | no |
(install-cpu)=
@ -230,6 +231,17 @@ JAX has experimental ROCm support. There are two ways to install JAX:
* Use [AMD's Docker container](https://hub.docker.com/r/rocm/jax); or
* Build from source (refer to {ref}`building-from-source` — a section called _Additional notes for building a ROCM `jaxlib` for AMD GPUs_).
(install-intel-gpu)=
## Intel GPU
Intel provides an experimental OneAPI plugin: intel-extension-for-openxla for Intel GPU hardware. For more details and installation instructions, refer to one of the following two methods:
1. Pip installation: [JAX acceleration on Intel GPU](https://github.com/intel/intel-extension-for-openxla/blob/main/docs/acc_jax.md).
2. Using [Intel's XLA Docker container](https://hub.docker.com/r/intel/intel-optimized-xla).
Please report any issues related to:
* JAX: [JAX issue tracker](https://github.com/jax-ml/jax/issues).
* Intel's OpenXLA plugin: [Intel-extension-for-openxla issue tracker](https://github.com/intel/intel-extension-for-openxla/issues).
## Conda (community-supported)
### Conda installation