Hi Annemarie

the computational complexity of the topology correction is quadratic in the (convex hull) of the largest defect. So small defects are fast and big ones take forever. However, if you have one that is that large it typically means that something else dramatic has gone wrong, like the hemispheres weren't properly separated, or the skull is attached to brain, or something like that. Take a look at the lh.orig.nofox or rh.orig.nofix and see if you can figure out what is wrong.

cheers Bruce

On Mon, 29 Jul 2013, Annemarie Brandhofe wrote:

Dear freesurfer community,

we are conducting cortical-thickness analysis using freesurfer-5.1.0 and have been experiencing some problems with the automatic recognition of the control points.  For several subjects, we run the recon-all command and after about 40 hours of computations, we receive the error message “Excessive topologic defect encountered: could not allocate 171060256 edges for retessellation

Cannot allocate memory”.  After the occurrence of the error message, the computation is stopped.

Below, we provide the details of our setup and the exact error message for one subject.

Feel  free to ask for more information. Thank you very much for your help!

Yours sincerely,

Annemarie

$ bugr

FREESURFER_HOME: /apps/prod/i486/freesurfer-5.1.0

Build stamp: freesurfer-Linux-centos4-stable-pub-v5.1.0

Debian version: 5.0.6

Kernel info: Linux 2.6.27.39-1-ql-beowulf i686

$ which freesurfer            ->          /apps/prod/i486/freesurfer-5.1.0/bin/freesurfer

Used command:

recon-all –autorecon2-cp –subjid IB34_mdeft.

Error message:

CORRECTING DEFECT 0 (vertices=72621, convex hull=6983)

Excessive topologic defect encountered: could not allocate 171060256 edges for retessellation

Cannot allocate memory

Linux beo-08 2.6.27.39-1-ql-beowulf #1 SMP Mon Nov 23 10:55:10 CET 2009 i686 GNU/Linux

recon-all -s IB34_mdeft. exited with ERRORS at Sat Jul 27 00:01:52 CEST 2013