[Eeglablist] speed of processing

Davida Streett davida_s2000 at yahoo.com
Thu Jun 1 05:24:49 PDT 2017

I appreciate this, thanks James.  I have been careful about setting the memory option to hold only one dataset.  
Yesterday I experimented with increasing the Java Heap Memory and while that seems to have helped a bit, its not enough of a change to make the full computation practical.  
At least during the Study>Plot Channel Measures processing, I see references to a very large array so I am intrigued by your suggestion of changing the way the results are accumulated. I'm not MATLAB-savvy and have just been depending on the EEGLAB gui, but will see if I can add your script; I think that might indeed help.
Thanks again!Davida  

      From: James Jones-Rounds <jj324 at cornell.edu>
 To: davida_s2000 at yahoo.com; EEGLAB List <eeglablist at sccn.ucsd.edu>; robert coben <drcoben at gmail.com> 
 Sent: Tuesday, May 30, 2017 3:35 PM
 Subject: Re: [Eeglablist] speed of processing
I've noticed these issues, sometimes, too. A couple things that have usually been the cause for me:
1) If you're working with a STUDY, have you made sure your Memory options are set to hold just one dataset in memory at a time? You can select this option from the File > "Memory and other options" menu.
2) Increase your MATLAB's Java Heap Memory to as much as you can (some other more savvy users out there might have some counter-recommendations about this). You can do this using the "Preferences > General > Java Heap Memory" menu option. You might need to restart MATLAB then for it to take effect.
3) Are you accumulating your results in any self-inflating matrix or array? i.e.
my_results = []
for data_index = 1:length(my_datasets)
       this_dataset_result = my_datasets{data_index}.results;       my_results = [my_results; this_dataset_result];

Sometimes, depending on how you do this, it can become excessively large and unweildy for your computer to continually store in memory.
I've also rarely found that the RAM usage reports from MATLAB correlate with how hard it seems my computer is working, so I don't know if it makes sense to trust the output you're talking about.
Hope that helps,
James Jones-RoundsLaboratory Manager
Human Development EEG and Psychophysiology (HEP) Laboratory,Department of Human Development,
Cornell University | Ithaca, NY
607-255-9883eeg at cornell.edu

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://sccn.ucsd.edu/pipermail/eeglablist/attachments/20170601/50ee7e21/attachment.html>

More information about the eeglablist mailing list