[Eeglablist] ICA time estimate

Robert Frank rmfrank at mac.com
Tue Jan 11 14:27:40 PST 2005


Dr. Kevin Glass of the Neuroinformatics Center at the University of 
Oregon has developed parallel C/C++ implementations of both  the 
FastICA and Infomax (Bell & Sejnowski) algorithms.

Parallel Infomax:

Our parallel Infomax implementation runs on up to 8 processors 
concurrently in  a shared memory configuration under Linux on an IBM 
p655/p690 cluster, or under AIX on a p655 (IBM Power Architecture).

Parallel FastICA:

Our parallel FastICA  implementation runs on up to 32 processors 
concurrently in a distributed memory configuration under Linux on a  
Dell Xeon cluster.

Both versions are designed to quickly decompose large, high density, 
EEG data sets.  If you were to send us the data, we can perform the ICA 
decomposition here on our machines and send you the results (Weight and 
Sphering matrices, independent components).

We have also developed sequential (single-processor) versions of both 
algorithms that you can run locally under Linux/OS X on an Intel or 
PowerPC machine, respectively.

These algorithms can be called directly or from within APECS (Automated 
Protocol for Electromagnetic Component Separation), a framework we 
developed that combines ICA decomposition with automatic ocular 
artifact detection and removal.

If you are interested in exploring either option, we would be happy top 
assist you.


Robert Frank
Neuroinformatics Center

On Jan 10, 2005, at 2:22 PM, Christian Bikle wrote:

> I started an ICA on a ~480meg data set which was acquired on a 
> 275channel system. The linux machine processing the ICA only has 1gig 
> of ram and a 2.8GHz processor. This process was running for over 2 
> weeks (I was on vacation during this time) until a co-worker shut down 
> the process (DOH!). I was wondering if any one had suggestions to 
> accelerate this process before starting again?
> Christian Bikle
> Psychologist
> Laboratory of Brain and Cognition
> National Institute of Mental Health
> Building 10/4C214
> Work: 3014518509
> _______________________________________________
> eeglablist mailing list eeglablist at sccn.ucsd.edu
> http://sccn.ucsd.edu/mailman/listinfo/eeglablist
> Eeglablist page: http://sccn.ucsd.edu/eeglab/eeglabmail.html
> To unsubscribe, send an empty email to 
> eeglablist-unsubscribe at sccn.ucsd.edu

More information about the eeglablist mailing list