On 08/19/2014 06:24 PM, Taylor, Marc D wrote:
I work with Alex. If you or anyone else can expound a bit more upon this statement on the website (http://surfer.nmr.mgh.harvard.edu/fswiki/Download):
" Important Note! When processing a group of subjects for your study, it is essential to process all your subjects with the same version of freesurfer, on the same OS platform and vendor, and for safety, even the same version of the OS. While we continue to work to ensure that results match across platforms (for instance the 32b and 64b CentOS 4 builds should produce identical results), there are none-the-less system-level libraries that are OS dependent. An exception to this rule is that you may view and edit files across any platform or version, and run some post-processing tools (outside the recon-all stream) if you check with us first (for instance you may run the longitudinal processing with newer versions)."
If I was being very simplistic, I would come away thinking that I should not even run OS patches on this system for the duration of the study. This has some implications for us if that is true.
Marc,
We make that recommendation because it removes effects related to changes in your platform from biasing the results or introducing a systematic error into the analysis. However, saying you should not apply OS patches is a too strict of an interpretation of what we are saying. Primarily, stick with the same major version release (e.g. RH6 RH7).
What about identical results across Linux vs Mac vs the Virtualbox? Is that not expected? What has been the experience of people using this software?
Definitely do not process some of your data on one OS and then the remaining data on a completely different one. This is exactly the kind of thing you want to avoid and could possibly introduce all types of biases in certain types of studies (e.g. longitudinal). Sticking to a the major release of a particular platform for an entire study is the best way to protect yourself against unintended secondary effects.
As I mentioned to Alex ina separate email, this may be more of an issue with Linux v. Mac than Linux v. Virtual Box, but the overall statement still applies.
Is there a listing of "some post-processing tools" that are known to work across all of the platforms? What about a list of the ones that are known not to work?
When you say "some post-processing tools" to you mean non-freesurfer related tools? If so, I can not provide input on that and you will have to investigate what those vendors/developers advertise. But if you mean something along the lines of "Is it ok to process results generated on one platform, say Linux, using post processing tools on another platform" I would say Yes, that is ok.
-Zeke
Thanks again for your time and consideration. Eagerly awaiting a response to this inquiry.
Marc Taylor
-----Original Message----- From: freesurfer-bounces@nmr.mgh.harvard.edu [mailto:freesurfer-bounces@nmr.mgh.harvard.edu] On Behalf Of Z K Sent: Tuesday, August 19, 2014 4:28 PM To: Freesurfer support list; Lazarevich, Alexander Subject: Re: [Freesurfer] FreeSurfer platform/performance questions
Hello Alex,
I created the FreeSurfer VM and you are correct that the Virtual Box host is 32bit. I cant remember exactly why I did it that way, but I suspect it may have been in an effort to reduce overall size, as the gzipped size of the .vdi file is 9gigs. Also, it was created more as matter of convenience to allow users with PCs the means for viewing and processing data, but it is not optimized for processing large scale data sets.
There is no 64bit VM currently available. However, it would not be too difficult to make one. Simply download any of the available 64bit host images (http://virtualboxes.org/images/), open it in VirtualBox, install 64bit freesurfer within the image, and then export the disk image which will now have freesurfer installed on it.
I could take on this task but I can not guarantee I will get to it in the immediate future.
-Zeke
On 08/18/2014 12:07 PM, Lazarevich, Alexander wrote:
One of the packages we have recently installed upon request is FreeSurfer. We have installed freesurfer-Linux-centos6_x86_64-stable-pub-v5.3.0.tar.gz ftp://surfer.nmr.mgh.harvard.edu/pub/dist/freesurfer/5.3.0/freesurfer -Linux-centos6_x86_64-stable-pub-v5.3.0.tar.gz on our Red Hat Enterprise Linux 6 workstations, as well as on our Windows 7 Enterprise 64bit workstations. We have also installed Oracle's VirtualBox on a Windows 7 Enterprise 64bit machine, and run the VM freesurfer-Virtualbox-linux-x86-stable-pub-v5.3-full.vdi.gz ftp://surfer.nmr.mgh.harvard.edu/pub/dist/freesurfer/5.3.0/freesurfer -Virtualbox-linux-x86-stable-pub-v5.3-full.vdi.gz inside that virtual environment.
The researchers who requested FreeSurfer are all part of the same group, and they are trying to complete a very large research project which requires FreeSurfer to process large amounts of data. They have normally been running FreeSurfer on the VM in the XUbuntu VirtualBox environment on their own average desktop PC's. However, they are concerned about their processing time with FreeSurfer, so they are hoping to take advantage of the Vislab computing resources to decrease the processing time in FreeSurfer.
We are running into the following problems/concerns:
1)The FreeSurfer VM (Xubuntu) that runs in VirtualBox is a 32bit host. Is there a 64bit VM available? That 32bit host can only access 4GB of memory, whereas they would like to be able to access more of the workstations 128GB of memory.
Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer
The information in this e-mail is intended only for the person to whom it is addressed. If you believe this e-mail was sent to you in error and the e-mail contains patient information, please contact the Partners Compliance HelpLine at http://www.partners.org/complianceline . If the e-mail was sent to you in error but does not contain patient information, please contact the sender and properly dispose of the e-mail.
Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer