.. _performance:

|
:Last modified: |today|
|

.. _perfrecommendations:

===========================
performance recommendations
===========================
#. Use :ref:`sge` to run jobs in parallel

    Grid Engine balances jobs across the cluster, so the jobs run more efficiently.

#. Avoid running more than 3 CPU-intensive jobs within your X2Go session

    X2Go servers have only 8 CPUs, and are used by many people. If you need to run jobs in parallel, use :ref:`sge`

#. Use only one server for both X2Go and secure shell (SSH) logins.

    Jobs started via secure shell run at low priority *unless* you log into the server where your X2Go session is running. 

#. Improve Jupyter Notebook performance by using a local web browser to :ref:`display the Jupyter notebook<jupyterperf>`.


#. Email me

    If you continue to experience performance problems, please email support-neuro@berkeley.edu. 

        Please include as much information as possible, e.g.
    
        * name of server, e.g. nx1
        * name of program/application 
        * command line used to run program
        * nature of problem, e.g. slow screen updates, analysis takes longer to complete, etc
        * error message