:orphan:

Training on CPUs
=========================



.. raw:: html

    <div class="sphx-glr-thumbnails">

.. thumbnail-parent-div-open

.. raw:: html

    <div class="sphx-glr-thumbcontainer" tooltip="Graph Neural Network (GNN) training suffers from low scalability on multi-core CPUs.  Specificially, the performance often caps at 16 cores, and no improvement is observed when applying more than 16 cores [#f1]_. ARGO is a runtime system that offers scalable performance.   With ARGO enabled, we are able to scale over 64 cores, allowing ARGO to speedup GNN training (in terms of epoch time) by up to 4.30x and 3.32x on a Xeon 8380H and a Xeon 6430L, respectively [#f2]_. This chapter focus on how to setup ARGO to unleash the potential of multi-core CPUs to speedup GNN training.">

.. only:: html

  .. image:: /tutorials/cpu/images/thumb/sphx_glr_argo_tutorial_thumb.png
    :alt:

  :ref:`sphx_glr_tutorials_cpu_argo_tutorial.py`

.. raw:: html

      <div class="sphx-glr-thumbnail-title">Improve Scalability on Multi-Core CPUs</div>
    </div>


.. raw:: html

    <div class="sphx-glr-thumbcontainer" tooltip="This chapter focus on providing best practises for environment setup to get the best performance during training and inference on the CPU.">

.. only:: html

  .. image:: /tutorials/cpu/images/thumb/sphx_glr_cpu_best_practises_thumb.png
    :alt:

  :ref:`sphx_glr_tutorials_cpu_cpu_best_practises.py`

.. raw:: html

      <div class="sphx-glr-thumbnail-title">CPU Best Practices</div>
    </div>


.. thumbnail-parent-div-close

.. raw:: html

    </div>


.. toctree::
   :hidden:

   /tutorials/cpu/argo_tutorial
   /tutorials/cpu/cpu_best_practises



.. only:: html

 .. rst-class:: sphx-glr-signature

    `Gallery generated by Sphinx-Gallery <https://sphinx-gallery.github.io>`_