# sgesvd returned error

**URL:** <https://mne.discourse.group/t/sgesvd-returned-error/475>\
**Category:** Mailing List Archive (read-only)\
**Tags:** list-archive\
**Created:** [August 17, 2012, 11:13pm UTC](https://mne.discourse.group/t/sgesvd-returned-error/475 "2012-08-17T23:13:21Z")\
**Posts on this page:** 3\
**Page:** 1

<div class="post-metadata">

**Author:** ![system](https://global.discourse-cdn.com/free1/uploads/mne/original/1X/85cc6bd2b69cb698a166dc6d880fb550510d0144.jpeg) [@system](https://mne.discourse.group/u/system)\
**Post date:** [August 17, 2012, 11:13pm UTC](https://mne.discourse.group/t/sgesvd-returned-error/475/1 "2012-08-17T23:13:21Z")

</div>

Dear experts,

For many datasets I was running "mne\_do\_inverse\_operator" and it worked  
fine. Today I hit a problem with the SVD of the weighted forward solution.

When I run the command with the same input with the eeg option everything  
works fine. But when I try to generate the MEG inverse operator I get the  
following error:

SVD.sgesvd returned error: 54

Do you have any ideas how to solve this problem?

Thanks,

Alex

-------------- next part --------------  
An HTML attachment was scrubbed...  
URL: [http://mail.nmr.mgh.harvard.edu/pipermail/mne\_analysis/attachments/20120817/6680df92/attachment.html](http://mail.nmr.mgh.harvard.edu/pipermail/mne_analysis/attachments/20120817/6680df92/attachment.html)

---

<div class="post-metadata">

**Author:** ![system](https://global.discourse-cdn.com/free1/uploads/mne/original/1X/85cc6bd2b69cb698a166dc6d880fb550510d0144.jpeg) [@system](https://mne.discourse.group/u/system)\
**Post date:** [August 19, 2012, 10:35am UTC](https://mne.discourse.group/t/sgesvd-returned-error/475/2 "2012-08-19T10:35:55Z")

</div>

Hi Alex,

Have you checked that your noise covariance contains reasonable values by loading it to Matlab or Python?

- Matti

---

<div class="post-metadata">

**Author:** ![system](https://global.discourse-cdn.com/free1/uploads/mne/original/1X/85cc6bd2b69cb698a166dc6d880fb550510d0144.jpeg) [@system](https://mne.discourse.group/u/system)\
**Post date:** [August 20, 2012, 2:39pm UTC](https://mne.discourse.group/t/sgesvd-returned-error/475/3 "2012-08-20T14:39:28Z")

</div>

Hi Matti,

Thanks for this hint. The data matrix of my current noise covariance  
contains from -2\*10^-10 to 6\*10^-10 which is much smaller than in an older  
matrix with values in the range of -9\*10^-6 to 2\*10^-5.

What really confuses me, is that it is working for the EEG inverse operator.  
Can you explain that?

Thanks,

Alex