[Eeglablist] Update on GPU and/or Parallel Processing Toolbox?

[James Desjardins]

Hi Joshua,

I have been working on the batch_context extension for eeglab that we hope will integrate eeglab with High Performance Computing clusters. Typically this extension submits jobs to cluster computer schedulers and allows for multiple parallel jobs (e.g. as you mentioned "for running multiple computations on different datasets in parallel").

There is a demonstration of the batch_context extension running MPI AMICA jobs on a SHARCNet cluster here: https://www.youtube.com/watch?v=32axMVYFFBA

The batch_context extension could easily be modified to run parpools and parfor loops on the local machines if that would be valuable. It is typically the case that this sort of "parallelization" is not beneficial on a multi-core machines as computations are typically optimized across the cores. That being said, if you notice that on an 8 core machine (for example) you are running a time consuming procedure that is bound to a single core and consumes less than one eighth of the machine's memory, then it would be valuable to run eight of these procedures at once. If this would be useful let me know and I will add it to the batch_context extension here: https://github.com/jadesjardins/batch_context.

I have not been using GPU optimization, but it sounds exciting!

James Desjardins, MA
High Performance Computing Consultant
SHARCNet, Compute/Calcul Canada
Brock University
500 Glenridge Ave.
St. Catharines, ON, Canada L2S 3A1
905-688-5550 x4676
--
________________________________
From: eeglablist-bounces at sccn.ucsd.edu [eeglablist-bounces at sccn.ucsd.edu] on behalf of Baker, Joshua [joshua.baker at ntu.ac.uk]
Sent: November 18, 2015 5:41 AM
To: eeglablist at sccn.ucsd.edu
Subject: [Eeglablist] Update on GPU and/or Parallel Processing Toolbox?


Dear EEGlab list



I see that on the EEGlab wiki, there have been some tests using the GPUmat toolbox (wiki updated in 2010). I have just purchased a new machine that i would wish to take full advantage of. Have there been any recent developments/tests on using either GPU computing for faster computations, or using the Parallel Processing Toolbox for running multiple computations on different datasets in parallel on the same machine?



Many thanks,







Joshua Baker. MSc.
Psychology Technician in Behavioural Neuroscience & PhD Researcher.
Midlands Representitive for the Association of Technical Staff in Psychology
Chaucer 433 (Psychology Lab Support)
Division of Psychology
Nottingham Trent University

Email:Joshua.baker at ntu.ac.uk<mailto:mJoshua.baker at ntu.ac.uk>
Tel: 01158 482017
Mob: 07791846148
Linkedin: https://uk.linkedin.com/pub/joshua-baker/27/626/815

DISCLAIMER: This email is intended solely for the addressee. It may contain private and confidential information. If you are not the intended addressee, please take no action based on it nor show a copy to anyone. In this case, please reply to this email to highlight the error. Opinions and information in this email that do not relate to the official business of Nottingham Trent University shall be understood as neither given nor endorsed by the University. Nottingham Trent University has taken steps to ensure that this email and any attachments are virus-free, but we do advise that the recipient should check that the email and its attachments are actually virus free. This is in keeping with good computing practice.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://sccn.ucsd.edu/pipermail/eeglablist/attachments/20151201/9abede9d/attachment.html>