SCIRun User Mailing List

Text archives Help


Re: [SCIRUN-USERS] additional question (memory allocation)


Chronological Thread 
  • From: David Weinstein <dmw@sci.utah.edu>
  • To: "David Himes" <dhimes@northstarneuro.com>
  • Cc: <scirun-users@sci.utah.edu>
  • Subject: Re: [SCIRUN-USERS] additional question (memory allocation)
  • Date: Mon, 10 May 2004 14:00:44 -0600
Hi David,

How large is the DWI volume that you're working with? BioTensor is notorious for using extra memory -- it computes a bunch of extra intermediate data in order to improve the interactivity of the visualization... but that can cause problems when the input data volume gets large. I'm happy to help you work around this.

As for 64-bit mode -- you can only enable that if you're running on a 64-bit architecture (e.g. SGI). If you are, you can build SCIRun/BioTensor from scratch with an --enable-64bit flag on the configure line. Please let me know if you require any assistance in getting that set up (if you're on a 64-bit machine) and I can help you.

Thanks for your interest in BioTensr -- I'm looking forward to getting you up and running with our software, and would be very interested in any feedback you could provide us on suggestions for further improving its usability.

Regards,
David Weinstein
NCRR Technical Manager
SCI Institute, University of Utah


On May 10, 2004, at 12:18 PM, David Himes wrote:

Gordon, assuming you're the one that monitors this list, I have, hopefully, one final question.  I've run BioTensor successfully with the gradient input advice you gave me.  The routine makes it all the way to the planes visualization step (just about all the way through) and then dies with a memory allocation error.  I have 512 MB of RAM and a 2 GB swap space.  I'm having a hard time imagining that all that memory has been consumed.  However, in browsing through your SCIRun FAQ's, I saw a question regarding a memory allocation error and it indicated that a 64-bit mode should be enabled.  I've searched exhaustively to see if there's any guidance to do this (assuming that this indeed, is the problem), but, unfortunately have come up empty.  I know that C++ routines are notorious for consuming memory and not freeing up processes.  Please advise.  Once again, thanks for all your help.  Please reply to this address.



_________________________
David Himes
Clinical Research Engineer

mailto:dhimes@northstarneuro.com

Direct line: 206.902.1928
Fax: 206.902.0140

Northstar Neuroscience, Inc.
2401 Fourth Avenue, Suite 300
Seattle, WA 98121

www.northstarneuro.com

Archive powered by MHonArc 2.6.16.

Top of page