[BUG] macstoolbox during stats step in parfor not working #1225
Labels
bug 🐛
Something isn't working
Comments
|
can you copy the script? just curious to know at what step this is happening |
|
Well, I guess we are witnessing again he magic of bug disappearing while having someone else attention remotely. I cannot reproduce the error and it is running now. Though it might be the case that first run in parallel was not able to copy something, it was able to it w/o parallel option, now running it in parallel does not complain anymore. I had a bad book keeping the first time and now I do not have the full error message. Are you still interested in the script? |
|
not any more!!! 😉 but if the shy bug build confidence and reappears, send it with the script. |
|
all right, not the same but within the same domain. running some stats step in parallel. this is the error: the error is printed n = jobs |
|
entire script is this one |
|
running without parfor works |
|
OK so I suspect that each worker in the parfor is trying to reinstall the MACs toolbox when initialising bidspm but it cannot because other workers are using it? |
|
I can try to make check the MACs toolbox version (or git commit shasum) before recopying it every time, this way it would only update it if it finds that the previous version is too old... |
|
Makes sense. No rush
…On Mon, 8 Apr 2024 at 17:47, Remi Gau ***@***.***> wrote:
I can try to make check the MACs toolbox version (or git commit shasum)
before recopying it every time, this way it would only update it if it
finds that the previous version is too old...
—
Reply to this email directly, view it on GitHub
<#1225 (comment)>,
or unsubscribe
<https://github.com/notifications/unsubscribe-auth/AJCWFPD5QHAN4NMQNDSOVOLY4K3X5AVCNFSM6AAAAABF4U5T6CVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDANBTGA4TQNBWHA>
.
You are receiving this because you authored the thread.Message ID:
***@***.***>
--
*Marco Barilari -- PhD student*
*twitter <https://twitter.com/marcobarilari> - github
<https://github.com/marcobarilari>*
Université Catholique de Louvain (UCL)
Psychological Sciences Research Institute (IPSY)
Institute of NeuroScience (IONS)
Place Cardinal Mercier, 10 - 1348 Louvain-la-Neuve
Desk: D.326
E-mail(1): marco.barilari [at] uclouvain.be ***@***.***>
E-mail(2): marcobarilr [at] gmail.com ***@***.***>
Website: www.cpplab.be
|
|
trying a fix in #1228 |
|
@marcobarilari |
|
Hi there, yes this indeed fixed the problem. However, a similar one happens when CPP-ROI copies the atlases. Here though is quite strange, it only happens if I run the script from terminal matlab call, while if I run the code (copy pasted) in the same terminal matlab session it work perfectly. in the file attached what is printed in the terminal/matlab command window 20240419_error_command_window_parallelpool_atalases_installation.txt |
|
I see that this is a specifc problem that I might be the only one having so no need to rush or find a solution if you think it might requires too much debugging. A workaround could be to let CPP ROI checking if the atlases are already there and skip copying in case. |
yup: can you open an issue in CPP_ROI and copy paste content from those 2 last comments there ? |
1 task
|
done, thanks |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Is there an existing issue for this?
Operating system
Operating system version
SPM 12 version
Platform
Platform version
bidspm version
v3.1.0
bidspm branch / commit number
branch =
commit =
Expected Behavior
No response
BIDS stats model
No response
Current Behavior & Error message
Anything else?
the erros message is repeated for all the jobs
The text was updated successfully, but these errors were encountered: