Dear Freesurfer-list,
I'm trying to construct an average group surface map using 'make_average_subject'. My group consists of a rather large number of subjects, around 400 subjects.
Unfortunately, running 'make_average_subject' gives an error when processing subject 253, saying it has an 'mgh reading error' on the 'T1.mgz'.
I checked that the involved files are all oke. In addition, I have run 'make_average_subject' several times with different group lists, but every time it stops at subject 253.
I'm using an Intel MacPro with 16 GB of system memory, and there is plenty of free memory when it runs.
Is there a maximum set on the number of subjects you can use in 'make_average_subject'? Can I reset this value?
Yours sincerely,
Martijn van den Heuvel
Hi Martijn,
it looks like we have a memory leak. We'll look into it and get back to you. I'm surprised you don't get past 253 with 16G of ram though. I've run it on 300 or so.
Bruce On Wed, 21 Feb 2007, martijn van den heuvel wrote:
Dear Freesurfer-list,
I'm trying to construct an average group surface map using 'make_average_subject'. My group consists of a rather large number of subjects, around 400 subjects.
Unfortunately, running 'make_average_subject' gives an error when processing subject 253, saying it has an 'mgh reading error' on the 'T1.mgz'.
I checked that the involved files are all oke. In addition, I have run 'make_average_subject' several times with different group lists, but every time it stops at subject 253.
I'm using an Intel MacPro with 16 GB of system memory, and there is plenty of free memory when it runs.
Is there a maximum set on the number of subjects you can use in 'make_average_subject'? Can I reset this value?
Yours sincerely,
Martijn van den Heuvel
Hi bruce,
Thanks for looking into my 'maximumnumber?' problem. Good to hear that you have run in on 300+, at least I know it should be possible and there is no fixed maximum number.
Hope to hear from you,
thanks, Martijn
On 2/21/07, Bruce Fischl fischl@nmr.mgh.harvard.edu wrote:
Hi Martijn,
it looks like we have a memory leak. We'll look into it and get back to you. I'm surprised you don't get past 253 with 16G of ram though. I've run it on 300 or so.
Bruce On Wed, 21 Feb 2007, martijn van den heuvel wrote:
Dear Freesurfer-list,
I'm trying to construct an average group surface map using 'make_average_subject'. My group consists of a rather large number of subjects, around 400 subjects.
Unfortunately, running 'make_average_subject' gives an error when
processing
subject 253, saying it has an 'mgh reading error' on the 'T1.mgz'.
I checked that the involved files are all oke. In addition, I have run 'make_average_subject' several times with different group lists, but
every
time it stops at subject 253.
I'm using an Intel MacPro with 16 GB of system memory, and there is
plenty
of free memory when it runs.
Is there a maximum set on the number of subjects you can use in 'make_average_subject'? Can I reset this value?
Yours sincerely,
Martijn van den Heuvel
sure, sorry for the bug. If you have a linux machine around with a fair amount of ram you should be able to run them though. Not sure why the mac runs out so much more quickly.
On Thu, 22 Feb 2007, martijn van den heuvel wrote:
Hi bruce,
Thanks for looking into my 'maximumnumber?' problem. Good to hear that you have run in on 300+, at least I know it should be possible and there is no fixed maximum number.
Hope to hear from you,
thanks, Martijn
On 2/21/07, Bruce Fischl fischl@nmr.mgh.harvard.edu wrote:
Hi Martijn,
it looks like we have a memory leak. We'll look into it and get back to you. I'm surprised you don't get past 253 with 16G of ram though. I've run it on 300 or so.
Bruce On Wed, 21 Feb 2007, martijn van den heuvel wrote:
Dear Freesurfer-list,
I'm trying to construct an average group surface map using 'make_average_subject'. My group consists of a rather large number of subjects, around 400 subjects.
Unfortunately, running 'make_average_subject' gives an error when
processing
subject 253, saying it has an 'mgh reading error' on the 'T1.mgz'.
I checked that the involved files are all oke. In addition, I have run 'make_average_subject' several times with different group lists, but
every
time it stops at subject 253.
I'm using an Intel MacPro with 16 GB of system memory, and there is
plenty
of free memory when it runs.
Is there a maximum set on the number of subjects you can use in 'make_average_subject'? Can I reset this value?
Yours sincerely,
Martijn van den Heuvel
thanks, Ill try to find a Linux machine and run the average.
martijn
On 2/22/07, Bruce Fischl fischl@nmr.mgh.harvard.edu wrote:
sure, sorry for the bug. If you have a linux machine around with a fair amount of ram you should be able to run them though. Not sure why the mac runs out so much more quickly.
On Thu, 22 Feb 2007, martijn van den heuvel wrote:
Hi bruce,
Thanks for looking into my 'maximumnumber?' problem. Good to hear that
you
have run in on 300+, at least I know it should be possible and there is
no
fixed maximum number.
Hope to hear from you,
thanks, Martijn
On 2/21/07, Bruce Fischl fischl@nmr.mgh.harvard.edu wrote:
Hi Martijn,
it looks like we have a memory leak. We'll look into it and get back to you. I'm surprised you don't get past 253 with 16G of ram though. I've run it on 300 or so.
Bruce On Wed, 21 Feb 2007, martijn van den heuvel wrote:
Dear Freesurfer-list,
I'm trying to construct an average group surface map using 'make_average_subject'. My group consists of a rather large number of subjects, around 400 subjects.
Unfortunately, running 'make_average_subject' gives an error when
processing
subject 253, saying it has an 'mgh reading error' on the 'T1.mgz'.
I checked that the involved files are all oke. In addition, I have
run
'make_average_subject' several times with different group lists, but
every
time it stops at subject 253.
I'm using an Intel MacPro with 16 GB of system memory, and there is
plenty
of free memory when it runs.
Is there a maximum set on the number of subjects you can use in 'make_average_subject'? Can I reset this value?
Yours sincerely,
Martijn van den Heuvel
let us know if it works, as it will change how high a priority fixing the bug is.
cheers, Bruce On Thu, 22 Feb 2007, martijn van den heuvel wrote:
thanks, Ill try to find a Linux machine and run the average.
martijn
On 2/22/07, Bruce Fischl fischl@nmr.mgh.harvard.edu wrote:
sure, sorry for the bug. If you have a linux machine around with a fair amount of ram you should be able to run them though. Not sure why the mac runs out so much more quickly.
On Thu, 22 Feb 2007, martijn van den heuvel wrote:
Hi bruce,
Thanks for looking into my 'maximumnumber?' problem. Good to hear that
you
have run in on 300+, at least I know it should be possible and there is
no
fixed maximum number.
Hope to hear from you,
thanks, Martijn
On 2/21/07, Bruce Fischl fischl@nmr.mgh.harvard.edu wrote:
Hi Martijn,
it looks like we have a memory leak. We'll look into it and get back to you. I'm surprised you don't get past 253 with 16G of ram though. I've run it on 300 or so.
Bruce On Wed, 21 Feb 2007, martijn van den heuvel wrote:
Dear Freesurfer-list,
I'm trying to construct an average group surface map using 'make_average_subject'. My group consists of a rather large number of subjects, around 400 subjects.
Unfortunately, running 'make_average_subject' gives an error when
processing
subject 253, saying it has an 'mgh reading error' on the 'T1.mgz'.
I checked that the involved files are all oke. In addition, I have
run
'make_average_subject' several times with different group lists, but
every
time it stops at subject 253.
I'm using an Intel MacPro with 16 GB of system memory, and there is
plenty
of free memory when it runs.
Is there a maximum set on the number of subjects you can use in 'make_average_subject'? Can I reset this value?
Yours sincerely,
Martijn van den Heuvel
Are you sure the problem isn't simply that, even though your machine is 64-bit, the FreeSurfer MacIntel distribution is only 32-bit? Cheers, Steve.
On 22 Feb 2007, at 15:00, martijn van den heuvel wrote:
thanks, Ill try to find a Linux machine and run the average.
martijn
On 2/22/07, Bruce Fischl < fischl@nmr.mgh.harvard.edu> wrote:sure, sorry for the bug. If you have a linux machine around with a fair amount of ram you should be able to run them though. Not sure why the mac runs out so much more quickly.
On Thu, 22 Feb 2007, martijn van den heuvel wrote:
Hi bruce,
Thanks for looking into my 'maximumnumber?' problem. Good to hear
that you
have run in on 300+, at least I know it should be possible and
there is no
fixed maximum number.
Hope to hear from you,
thanks, Martijn
On 2/21/07, Bruce Fischl fischl@nmr.mgh.harvard.edu wrote:
Hi Martijn,
it looks like we have a memory leak. We'll look into it and get
back to
you. I'm surprised you don't get past 253 with 16G of ram
though. I've
run it on 300 or so.
Bruce On Wed, 21 Feb 2007, martijn van den heuvel wrote:
Dear Freesurfer-list,
I'm trying to construct an average group surface map using 'make_average_subject'. My group consists of a rather large
number of
subjects, around 400 subjects.
Unfortunately, running 'make_average_subject' gives an error when
processing
subject 253, saying it has an 'mgh reading error' on the '
T1.mgz'.
I checked that the involved files are all oke. In addition, I
have run
'make_average_subject' several times with different group
lists, but
every
time it stops at subject 253.
I'm using an Intel MacPro with 16 GB of system memory, and
there is
plenty
of free memory when it runs.
Is there a maximum set on the number of subjects you can use in 'make_average_subject'? Can I reset this value?
Yours sincerely,
Martijn van den Heuvel
Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer
------------------------------------------------------------------------ --- Stephen M. Smith, Professor of Biomedical Engineering Associate Director, Oxford University FMRIB Centre
FMRIB, JR Hospital, Headington, Oxford OX3 9DU, UK +44 (0) 1865 222726 (fax 222717) steve@fmrib.ox.ac.uk http://www.fmrib.ox.ac.uk/~steve ------------------------------------------------------------------------ ---
Martijn,
The problem you are encountering with the number of subjects allowed by make_average_subject does not appear to be a memory leak as originally thought, but rather simply a 'too many open files' limit. On the Mac, the limit to the number of open files is 256 (on Linux its 1024).
The solution is simple, type this in your environment, or include it near the top of your make_average_subjects script:
limit descriptors 2048
This ups the limit to 2048 files (supporting # of subjects-3).
Type 'limit' to confirm that the setting has changed. Then run make_average_subject in that window.
This fix will appear in the next release.
Nick
On Wed, 2007-02-21 at 11:11 -0500, Bruce Fischl wrote:
Hi Martijn,
it looks like we have a memory leak. We'll look into it and get back to you. I'm surprised you don't get past 253 with 16G of ram though. I've run it on 300 or so.
Bruce On Wed, 21 Feb 2007, martijn van den heuvel wrote:
Dear Freesurfer-list,
I'm trying to construct an average group surface map using 'make_average_subject'. My group consists of a rather large number of subjects, around 400 subjects.
Unfortunately, running 'make_average_subject' gives an error when processing subject 253, saying it has an 'mgh reading error' on the 'T1.mgz'.
I checked that the involved files are all oke. In addition, I have run 'make_average_subject' several times with different group lists, but every time it stops at subject 253.
I'm using an Intel MacPro with 16 GB of system memory, and there is plenty of free memory when it runs.
Is there a maximum set on the number of subjects you can use in 'make_average_subject'? Can I reset this value?
Yours sincerely,
Martijn van den Heuvel
Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer
freesurfer@nmr.mgh.harvard.edu
-
Bruce Fischl -
martijn van den heuvel -
Nick Schmansky -
Steve Smith