[Eeglablist] Memory issue using Clean Rawdata and ASR

[Cedric Cannard]

Wow Arno! That sounds unimaginably fast :)

James, I believe 64 = 64 MB. You may also use what's currently available instead of a specific amount, as Arno said in a previous email, the defined amount is to facilitate replication of findings (by defining the block size). Here is an example to use 85% of available memory from all arrays:

cutoff = 20;
m = memory;
usemem = .85;    % 85% of all available memory
maxmem = usemem * (m.MemAvailableAllArrays/1000000);  
disp(['Using ' num2str(usemem*100) '% of available memory (' num2str(round(maxmem/1000,1)) ' GB)'])
EEG = clean_asr(EEG,cutoff,[],[],[],[],[],[],[],[],maxmem);

And you can also use GPU if you are doing offline processing and have a good graphics card and the parallel toolbox. Simply add 1 after the maxmem input. 

However, note that this does not provide the option to remove sections like clean_rawdata does, it only reconstructs which can be problematic before ICA. But you can find the code to do that in clean_artifact just after the function is called line 273:283 

Cheers,

Cedric Cannard



------- Original Message -------
On Friday, September 30th, 2022 at 4:56 PM, Delorme, Arnaud via eeglablist <eeglablist at sccn.ucsd.edu> wrote:


> There is no maximum except MATLAB array size as long as you have the RAM.
> The maximum I was ever able to reserve was 1Tb of RAM on one of the large memory nodes of the San Diego supercomputer (var = rand(5000,5000, 5000); just bragging here :-).
> 
> Arno
> 
> _______________________________________________
> Eeglablist page: http://sccn.ucsd.edu/eeglab/eeglabmail.html
> To unsubscribe, send an empty email to eeglablist-unsubscribe at sccn.ucsd.edu
> For digest mode, send an email with the subject "set digest mime" to eeglablist-request at sccn.ucsd.edu