[Freesurfer] error with mri_ca

Hi Nick,

I keep getting the same error with mri_ca_register. I have now tried this with two different subjects and they both fail at about the same stage. I have checked both brainmask.mgz files in tkmedit for any skull stripping or talairach registration errors and they look fine. Like you had recommended, I have also checked the messages in the /var/log directory but found nothing to suggest that the system was running out of memory. I'm fresh out of ideas. Please let me know if there's any other thing that comes to mind.
I've attached the logs again for reference.

Binyam

Binyam Nardos
Cognitive Rehabilitation Research Group
Washington University School of Medicine Box # 8505
4444 Forest Park Blvd
St. Louis, MO 63108
Office: (314) 454-7795
Fax: (314) 286-1601

The contents of this e-mail message and any attachments are intended solely for the addressee(s) named in this message. This communication is intended to be and to remain confidential and may be subject to applicable physician/patient and/or work product privileges. If you are not the intended recipient of this message, or if this message has been addressed to you in error, please immediately alert the sender by reply e-mail and then delete this message and its attachments.

-----Original Message-----
From: Nick Schmansky [mailto:nicks@nmr.mgh.harvard.edu]
Sent: Thu 2/8/2007 3:31 PM
To: Binyam Nardos
Subject: RE: [Freesurfer] error with mri_ca_register

Binyam,

My run of mri_ca_register using the stable rh9 build of freesurfer also
completed successfully, so I was not able to recreate the segfault
problem you are seeing.

One thing I just learned is that you can view the file:

/var/log/messages

and check it for messages that say "Out of Memory". This sort of message
would appear if a freesurfer binary (or any other) is unable to allocate
memory. You will need root or sudo access to read the file on your
system.

Nick

On Wed, 2007-02-07 at 16:46 -0600, Binyam Nardos wrote:
> Nick,
> upgrading my RAM is definitely on top of my list. As far as the
> documentation that I mentioned, I don't think it's something you can
> update. I just remembered that I had seen that on the
> Dale/Fischl/Sereno 1999 paper (Cortical Surface-Based Analysis I),
> which, I know is pretty old, vis-a-vis all the time-consuming changes
> that I'm sure have been implemented since then. In any case, I think
> the paper reads "The entire procedure, from intensity normalization,
> through segmentation, tessellation, and inflation, runs completely
> automatically in about 1.5h on current PC architectures ..."
>
> Binyam
>
> Binyam Nardos
> Cognitive Rehabilitation Research Group
> Washington University School of Medicine Box # 8505
> 4444 Forest Park Blvd
> St. Louis, MO 63108
> Office: (314) 454-7795
> Fax: (314) 286-1601
>
>
>
> The contents of this e-mail message and any attachments are intended solely for the addressee(s) named in this message. This communication is intended to be and to remain confidential and may be subject to applicable physician/patient and/or work product privileges. If you are not the intended recipient of this message, or if this message has been addressed to you in error, please immediately alert the sender by reply e-mail and then delete this message and its attachments.
>
>
> ______________________________________________________________________
> From: Nick Schmansky [mailto:nicks@nmr.mgh.harvard.edu]
> Sent: Wed 2/7/2007 4:35 PM
> To: Binyam Nardos
> Subject: RE: [Freesurfer] error with mri_ca_register
>
>
>
> The mri_ca_register step typically takes about 10 hours to complete.
> It
> took 9 hours 37 minutes on the 2GHz machine I ran it on yesterday.
>
> Also, 1GB of RAM is pretty low. All of our machines have at least 2GB
> (our minimum recommended), most 4GB. This might actually be the
> source
> of your problem, as our code does not check when it runs out of memory
> (very bad, I know). I'll know more when this second run completes.
> But
> if it is at all possible to upgrade to 2GB or 4GB of RAM, I *highly*
> recommend it. I know with certainty that when someone on the mailing
> list upgraded from 512MB to 2GB, these sorts of problems disappeared.
>
> Can you point me to the documentation that says the segmentation takes
> 2-3 hours to complete? I'd like to update that.
>
>
> On Wed, 2007-02-07 at 16:20 -0600, Binyam Nardos wrote:
> > thanks Nick,
> > I tried processing another subject yesterday and again got stuck
> > somewhere around mri_ca_register, I believe it said (segmentation
> > error ... sorry I dont' have the details because I'm in windows
> world
> > right now.) In any case, I had another quick question for you:
> > I am running Freesurfer on ubuntu, which is installed on a separate
> HD
> > on my computer. Some of my specs are 1GB RAM, 3.2GHz dual processor.
> > In some of the documetation that I've read, a typical pc takes about
> > 2-3 hours for complete segmentation. It's been taking me about 3
> times
> > that (i'm actually yet to fully segment a subject successfuly ). But
> > is this normal?
> >
> > thanks again,
> > Binyam
> >
> > Binyam Nardos
> > Cognitive Rehabilitation Research Group
> > Washington University School of Medicine Box # 8505
> > 4444 Forest Park Blvd
> > St. Louis, MO 63108
> > Office: (314) 454-7795
> > Fax: (314) 286-1601
> >
> >
> >
> > The contents of this e-mail message and any attachments are intended
> solely for the addressee(s) named in this message. This communication
> is intended to be and to remain confidential and may be subject to
> applicable physician/patient and/or work product privileges. If you
> are not the intended recipient of this message, or if this message has
> been addressed to you in error, please immediately alert the sender by
> reply e-mail and then delete this message and its attachments.
> >
> >
> >
> ______________________________________________________________________
> > From: Nick Schmansky [mailto:nicks@nmr.mgh.harvard.edu]
> > Sent: Wed 2/7/2007 4:10 PM
> > To: Binyam Nardos
> > Subject: RE: [Freesurfer] error with mri_ca_register
> >
> >
> >
> > Binyam,
> >
> > I was able to run mri_ca_register successfully with this subject,
> but
> > I
> > was using a 'development' build (not the one used publicly), and I
> was
> > running on a 64bit system. So this problem is not a blatant bug in
> > the
> > code.
> >
> > I am going to re-run mri_ca_register again, this time using the rh9
> > stable build (which is what you used) to see if I can get it to
> > segfault
> > (the problem that you are encountering).
> >
> > Nick
> >
> >
> > On Tue, 2007-02-06 at 12:01 -0600, Binyam Nardos wrote:
> > > thanks Nick
> > >
> > > Binyam
> > >
> > > Binyam Nardos
> > > Cognitive Rehabilitation Research Group
> > > Washington University School of Medicine Box # 8505
> > > 4444 Forest Park Blvd
> > > St. Louis, MO 63108
> > > Office: (314) 454-7795
> > > Fax: (314) 286-1601
> > >
> > >
> > >
> > > The contents of this e-mail message and any attachments are
> intended
> > > solely for the addressee(s) named in this message. This
> > communication
> > > is intended to be and to remain confidential and may be subject to
> > > applicable physician/patient and/or work product privileges. If
> you
> > > are not the intended recipient of this message, or if this message
> > has
> > > been addressed to you in error, please immediately alert the
> sender
> > by
> > > reply e-mail and then delete this message and its attachments.
> > >
> > >
> > >
> > > -----Original Message-----
> > > From: Nick Schmansky [mailto:nicks@nmr.mgh.harvard.edu]
> > > Sent: Tue 2/6/2007 11:41 AM
> > > To: Binyam Nardos
> > > Subject: RE: [Freesurfer] error with mri_ca_register
> > >
> > > Binyam
> > >
> > > I've received the files and will have a look at them today.
> > >
> > > Nick
> > >
> > >
> > > On Tue, 2007-02-06 at 11:32 -0600, Binyam Nardos wrote:
> > > > Hi Nick,
> > > > I have just uploaded the 4 mgz files for the subject. I hope
> > that's
> > > > what you were looking for. I checked the skull-strip and
> talairach
> > > > registration and it looks fine.
> > > >
> > > >
> > > >
> > > > thanks,
> > > > Binyam
> > > >
> > > > Binyam Nardos
> > > > Cognitive Rehabilitation Research Group
> > > > Washington University School of Medicine Box # 8505
> > > > 4444 Forest Park Blvd
> > > > St. Louis, MO 63108
> > > > Office: (314) 454-7795
> > > > Fax: (314) 286-1601
> > > >
> > > >
> > > >
> > > > The contents of this e-mail message and any attachments are
> > intended
> > > > solely for the addressee(s) named in this message. This
> > > communication
> > > > is intended to be and to remain confidential and may be subject
> to
> > > > applicable physician/patient and/or work product privileges. If
> > you
> > > > are not the intended recipient of this message, or if this
> message
> > > has
> > > > been addressed to you in error, please immediately alert the
> > sender
> > > by
> > > > reply e-mail and then delete this message and its attachments.
> > > >
> > > >
> > > >
> > > > -----Original Message-----
> > > > From: Nick Schmansky [mailto:nicks@nmr.mgh.harvard.edu]
> > > > Sent: Tue 2/6/2007 11:18 AM
> > > > To: Binyam Nardos
> > > > Cc: Freesurfer Mailing List
> > > > Subject: Re: [Freesurfer] error with mri_ca_register
> > > >
> > > > Binyam,
> > > >
> > > > The recon-all.log indicates that mri_ca_register hit a
> > segmentation
> > > > fault, which is bad. Can you send me the subject data so that I
> > can
> > > > try
> > > > to recreate this segfault in a debugger? You can upload to me
> via
> > > the
> > > > filedrop:
> > > >
> > > > https://www.nmr.mgh.harvard.edu/facility/filedrop/index.html
> > > >
> > > > In the meantime, also look at the brainmask.mgz file with
> tkmedit
> > > and
> > > > confirm that it is skull-stripped and that the cerebellum is
> > > attached.
> > > > Also check the talairach registration with tkregister2.
> > > >
> > > > Nick
> > > >
> > > >
> > > > On Tue, 2007-02-06 at 10:39 -0600, Binyam Nardos wrote:
> > > > > hi all,
> > > > > I just ran autorecon2 on one of our subjects (for the second
> > time,
> > > > > after getting a similar error, i.e. ERROR: mri_ca_register
> with
> > > non-
> > > > > zero status. I just rerun recon-all with -autorecon2 without
> > > > > removing/changing anything produced by the previous
> autorecon2).
> > > Is
> > > > > there any logical place to start to fix the problem or do I
> have
> > > to
> > > > > rerun mri_ca_register as shown below:
> > > > >
> > > > > mri_ca_register -align -nobigventricles -mask brainmask.mgz -T
> > > > > transforms/talairach.lta
> > > > > norm.mgz /usr/local/freesurfer//average/RB_all_2006-02-15.gca
> > > > > transforms/talairach.m3z
> > > > >
> > > > > I have attached the recon-all log for reference.
> > > > > Please advise,
> > > > >
> > > > > thanks,
> > > > > Binyam
> > > > >
> > > > >
> > > > > Binyam Nardos
> > > > > Cognitive Rehabilitation Research Group
> > > > > Washington University School of Medicine Box # 8505
> > > > > 4444 Forest Park Blvd
> > > > > St. Louis, MO 63108
> > > > > Office: (314) 454-7795
> > > > > Fax: (314) 286-1601
> > > > >
> > > > >
> > > > >
> > > > > The contents of this e-mail message and any attachments are
> > > intended
> > > > > solely for the addressee(s) named in this message. This
> > > > communication
> > > > > is intended to be and to remain confidential and may be
> subject
> > to
> > > > > applicable physician/patient and/or work product privileges.
> If
> > > you
> > > > > are not the intended recipient of this message, or if this
> > message
> > > > has
> > > > > been addressed to you in error, please immediately alert the
> > > sender
> > > > by
> > > > > reply e-mail and then delete this message and its attachments.
> > > > >
> > > > >
> > > > >
> > > > >
> > > > > _______________________________________________
> > > > > Freesurfer mailing list
> > > > > Freesurfer@nmr.mgh.harvard.edu
> > > > > https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer
> > > >
> > > >
> > > >
> > >
> > >
> > >
> >
> >
> >
>
>
>