Re: [Freesurfer] mri_ca_label with high resolution data

Hi Falk

yes, it's an active area of devlopment for us cheers Bruce

On Thu, 21 Jul 2016, Falk Lüsebrink wrote:

Hi Bruce,

I thought so. Are there any other means to accelerate this procedure (in the future)?

Best, Falk

-----Ursprüngliche Nachricht----- Von: Bruce Fischl [mailto:fischl@nmr.mgh.harvard.edu] Gesendet: Donnerstag, 21. Juli 2016 15:52 An: Falk Lüsebrink Cc: Freesurfer support list Betreff: Re: AW: AW: [Freesurfer] mri_ca_label with high resolution data

Hi Falk

not every part of it is accelerated, but some are

cheers Bruce

On Thu, 21 Jul 2016, Falk Lüsebrink wrote:

...

Hi Bruce,

I'm using the current dev build:

ProgramName: mri_ca_label ProgramArguments: --all-info ProgramVersion: $Name: $ TimeStamp: 2016/07/21-13:42:39-GMT BuildTimeStamp: Jul 12 2016 13:31:56 CVS: $Id: mri_ca_label.c,v 1.113 2016/05/13 18:02:49 fischl Exp $ User: luesebrink Machine: reco Platform: Linux PlatformVersion: 3.16.0-4-amd64 CompilerName: GCC CompilerVersion: 40400

and in the log it says to use 12 threads

== Number of threads available to mri_ca_label for OpenMP = 12 ==

however, at least while saving the intensity scale

saving intensity scales to aseg.auto_noCCseg.label_intensities.txt saving sequentially combined intensity scales to aseg.auto_noCCseg.label_intensities.txt

it doesn't seem to help.

Best, Falk

---

Von: Bruce Fischl [fischl@nmr.mgh.harvard.edu] Gesendet: Donnerstag, 21. Juli 2016 15:30 An: Falk Lüsebrink Cc: Freesurfer support list Betreff: Re: AW: [Freesurfer] mri_ca_label with high resolution data

Hi Falk

the dev version of mri_ca_label does use openmp, so should be faster.

cheers Bruce On Thu, 21 Jul 2016, Falk Lüsebrink wrote:

...

Hi Bruce,

I started the process again with openmp set to 12 using the dev build from 12th July. However, mri_ca_label uses only 1 thread at that point. So I don't assume any faster processing.

Best, Falk

-----Ursprüngliche Nachricht----- Von: freesurfer-bounces@nmr.mgh.harvard.edu [mailto:freesurfer-bounces@nmr.mgh.harvard.edu] Im Auftrag von Bruce Fischl Gesendet: Dienstag, 19. Juli 2016 15:37 An: Freesurfer support list Betreff: Re: [Freesurfer] mri_ca_label with high resolution data

Hi Falk

I think openmp will help with this in V6, but probably not before Bruce

On Tue, 19 Jul 2016, Falk Lüsebrink wrote:

...

Hi all,

I'm currently processing a 250 µm MPRAGE with a dev build from mid of May. Running recon-all with default parameters ran flawlessly. Afterwards I added the -hires flag to process the data without conformation. However, since Saturday morning it is kind of stuck at mri_ca_label stating:

saving sequentially combined intensity scales to aseg.auto_noCCseg.label_intensities.txt

while consuming around 30 GB of memory. I ran something similar before and processing works, it's just terribly slow. Maybe this particular stage is just not very efficient as it does take a few minutes on 1 mm data only? Would it help to add the openmp flag in that case?

Best,

Falk

---

Falk Lüsebrink, M. Sc.

Otto-von-Guericke-Universität Magdeburg

Forschungscampus STIMULATE

http://www.forschungscampus-stimulate.de/

Universitätsplatz 2

39106 Magdeburg

Raum: ExFa - 4.06

Telefon: 0391-67-19366

Fax: 0391-67-19347

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.