Hi Nárlon,

Unfortunately, this is a known bug with parallel processing in FS 7.1.1. A fix will make it into the next v7 patch. As you mentioned, this error occurs at random, so in the meantime, you can mitigate this by either:

    - Rerunning recons until they don’t fail (not recommended)

    - Rerunning these failing recons without the parallel flag enabled

    - Patching your version of recon-all by commenting out or removing the following (line 4217) from $FREESURFER_HOME/bin/recon-all:

                if($OMP_NUM_THREADS > 1) set cmd = ($cmd --parallel)

Hope that helps,

Andrew

From: <freesurfer-bounces@nmr.mgh.harvard.edu> on behalf of "Boa Sorte Silva, Narlon" <narlon.silva@ubc.ca>
Reply-To: FS Help <freesurfer@nmr.mgh.harvard.edu>
Date: Monday, December 14, 2020 at 7:31 PM
To: FS Help <freesurfer@nmr.mgh.harvard.edu>
Subject: [Freesurfer] Recon-all error: MatrixMultiply: m1 is null!

        External Email - Use Caution        

Hi there,

Hope this finds you well

I have noticed the following error for some of my runs:

@# Cortical ribbon mask Mon Dec 14 15:43:24 PST 2020

/Applications/freesurfer/7.1.1/subjects/BT_long/BT_055_base/mri

\n mris_volmask --aseg_name aseg.presurf --label_left_white 2 --label_left_ribbon 3 --label_right_white 41 --label_right_ribbon 42 --save_ribbon --parallel BT_055_base \n

SUBJECTS_DIR is /Applications/freesurfer/7.1.1/subjects/BT_long

loading input data...

Running hemis in parallel

Processing left hemi

Processing right hemi

computing distance to right white surface 

computing distance to left white surface 

error: MatrixMultiply: m1 is null!

Darwin PHTH-CDM-NBOASORT 18.5.0 Darwin Kernel Version 18.5.0: Mon Mar 11 20:40:32 PDT 2019; root:xnu-4903.251.3~3/RELEASE_X86_64 x86_64

recon-all -s BT_055_base exited with ERRORS at Mon Dec 14 15:43:27 PST 2020

Here are details of the computer and FS version I am using:

• OS: macOS Mojave (10.14.4)
• FS: 7.1.1

It seems to happen at random, whenever I run recon-all on the same subject again, the error doesn’t occur. This is true for cross-sectional, base, and long processing steps in my machine. I have seen an email relating to this issue before, but I wasn’t able to access the attachment file which would fix the problem.

I was wondering if you guys would be able to help?

Thank you

Nárlon Cássio

Nárlon Cássio Boa Sorte Silva, PhD
CIHR/MSFHR Postdoctoral Research Fellow

Aging, Mobility, and Cognitive Neuroscience Lab

Djavad Mowafaghian Centre for Brain Health
University of British Columbia

Twitter: @BoaNarlon