Dear Xavier,

I am happy to hear you are using our examples :)

The tutorials are intentionally kept simple and therefore do not support MPI. While NEST’s internal
functions do support MPI, data retrieval operations (such as `get("events")` or `GetConnections()`)
are executed independently on each MPI process. The variations in the error messages occur because
different settings lead to different distributions of neurons across MPI processes and thus also
process-specific neuron numbers and behavior.

By default, recorders store their data in memory. The simplest way to collect results is to save
them to disk by adding the parameter `"record_to": "ascii"` to the recorder settings. This will
write the data to files. After running the simulation, you can load the results on a single MPI
process (e.g., using `if nest.Rank() == 0`), and the resulting matrix should then have the correct
shape.

The line

```python
nest.GetConnections(nrns_rec[0], nrns_rec[1:3]).set([params_init_optimizer] * 2)
```

is included only as a demonstration that dynamic variables can be initialized manually. In the
tutorial, they are set to their default values, so if you are not changing them explicitly, you can
safely comment out this line.

Further adjustments to the code may be necessary following these principles.

Hopefully this helps!

All the best,
Agnes

On 2025-10-29 17:12, Xavier Otazu Porter wrote:
> Hi Camilo,
>
> Thanks a lot for your help ... but problem remains. Now I get:
>
> Traceback (most recent call last):
>   File "eprop_supervised_regression_handwriting_bsshslm_2020.py", line 545, in <module>
>     readout_signal = readout_signal.reshape((readout_signal[0], n_iter, batch_size,
> steps["sequence"])) #readout_signal = readout_signal.reshape((n_out, n_iter, batch_size,
> steps["sequence"]))
> TypeError: only integer scalar arrays can be converted to a scalar index
>
> Xavier
> ----------------------------------------------------------------------------------------------------
> *From:* Camilo Jara Do Nascimento <C.Jara-Do-Nascimento@sussex.ac.uk>
> *Sent:* Wednesday, October 29, 2025 4:56 PM
> *To:* NEST User Mailing List <users@nest-simulator.org>; Xavier Otazu Porter <xotazu@cvc.uab.cat>
> *Subject:* Re: [NEST Users] MPI problem with e-prop tutorials
>  
> Hi Xavier,
>
> The problem I see is that the readout first dimension is based on the n_out, and when using mpi to
> parallelizing it will use this first dimension to parallelize your instances. I quickly fixed it
> by changing the first dimension to be readout_signal.shape[0] since its already defined before by
> using the "idc" indexes. 
>
> readout_signal = readout_signal.reshape((readout_signal.shape[0], n_iter,
> batch_size, steps["sequence"]))
>
> Hope this helps.
>
> Best,
> Camilo Jara D.
> ----------------------------------------------------------------------------------------------------
> *From:* Xavier Otazu <xotazu@cvc.uab.cat>
> *Sent:* 29 October 2025 14:03
> *To:* users@nest-simulator.org <users@nest-simulator.org>
> *Subject:* [NEST Users] MPI problem with e-prop tutorials
>  
> Hello,
>
> I have tried to execute one of the e-prop tutorials
> (eprop_supervised_regression_handwriting_bsshslm_2020.py) and get into problems when activating
> MPI. I do not know whether it is an e-prop C++ implementation problem or a Python tutorial code
> problem (in some cases, but not all of them, it looks like the last one).
>
> I have two files (hostfile_orig_1 and  hostfile_orig_2) to define in what nodes of the cluster to
> execute the program. The content of these files is:
> hostfile_orig_1:
> node0 slots=1
> node1 slots=1
>
> hostfile_orig_2:
> node0 slots=2
> node0 slots=2
>
> The 'slots' key tells how many (mpi) processes can be executed on a particular node.
>
> Depending on the number of processes, the errors are a bit different. In all the following
> examples, I use "total_num_virtual_procs": 2 on line 173 of tutorial file.
>
> Below I write the execution command and the output error.
>
> Command: mpirun -np 1 -hostfile hostfile_orig_1 python3
> eprop_supervised_regression_handwriting_bsshslm_2020.py
> Output: With a "serial" execution like this one, everything is OK:
>
> Command: mpirun -np 2 -hostfile hostfile_orig_2 python3
> eprop_supervised_regression_handwriting_bsshslm_2020.py
> Output:
> Traceback (most recent call last):
>   File "eprop_supervised_regression_handwriting_bsshslm_2020.py", line 404, in <module>
>     nest.GetConnections(nrns_rec[0], nrns_rec[1:3]).set([params_init_optimizer] * 2)
>   File "/home/neurobit/local/nest_3.9/lib64/python3.8/site-packages/nest/lib/hl_api_types.py",
> line 945, in set
>     raise TypeError("status dict must be a dict, or a list of dicts of length
> {}".format(self.__len__()))
> TypeError: status dict must be a dict, or a list of dicts of length 1
>
> Command: mpirun -np 2 -hostfile hostfile_orig_2 python3
> eprop_supervised_regression_handwriting_bsshslm_2020.py (commenting line 404)
> Output:
> Traceback (most recent call last):
>   File "eprop_supervised_regression_handwriting_bsshslm_2020.py", line 545, in <module>
>     readout_signal = readout_signal.reshape((n_out, n_iter, batch_size, steps["sequence"]))
> ValueError: cannot reshape array of size 364800 into shape (2,200,1,1824)
>
> Command: mpirun -np 2 -hostfile hostfile_orig_1 python3
> eprop_supervised_regression_handwriting_bsshslm_2020.py (commenting line 404)
> Output:
> Traceback (most recent call last):
>   File "eprop_supervised_regression_handwriting_bsshslm_2020.py", line 545, in <module>
>     readout_signal = readout_signal.reshape((n_out, n_iter, batch_size, steps["sequence"]))
> ValueError: cannot reshape array of size 364800 into shape (2,200,1,1824)
>
> Command: mpirun -np 4 -hostfile hostfile_orig_2 python3
> eprop_supervised_regression_handwriting_bsshslm_2020.py (commenting line 404 and using
> "total_num_virtual_procs": 4 on line 173 of tutorial file.)
> Output:
> Traceback (most recent call last):
>   File "eprop_supervised_regression_handwriting_bsshslm_2020.py", line 493, in <module>
>     "rec_out": get_weights(nrns_rec, nrns_out),
>   File "eprop_supervised_regression_handwriting_bsshslm_2020.py", line 482, in get_weights
>     conns["senders"] = np.array(conns["source"]) - np.min(conns["source"])
> TypeError: tuple indices must be integers or slices, not str
>
> In this last case, program stops and hungs.
>
> If you want, I could submit a bug report on github.
>
> Xavier
> _______________________________________________
> NEST Users mailing list -- users@nest-simulator.org
> To unsubscribe send an email to users-leave@nest-simulator.org
>
>
> Computer Vision Center <http://www.cvc.uab.cat>
> CONFIDENTIALITY WARNING <http://www.cvc.uab.es/?page_id=7475>
>
>
> _______________________________________________
> NEST Users mailing list -- users@nest-simulator.org
> To unsubscribe send an email to users-leave@nest-simulator.org

--
Agnes Korcsak-Gorzo
Institute for Advanced Simulation (IAS-6)       Phone: +49 2461/61-4748
Computational and Systems Neuroscience          Fax: +49 2461/61-9460
Jülich Research Centre,
Jülich, Germany

https://www.fz-juelich.de/en/ias/ias-6

Forschungszentrum Jülich GmbH
52425 Jülich
Sitz der Gesellschaft: Jülich
Eingetragen im Handelsregister des Amtsgerichts Düren Nr. HR B 3498
Vorsitzender des Aufsichtsrats: MinDir Stefan Müller
Geschäftsführung: Prof. Dr. Astrid Lambrecht (Vorsitzende),
Dr. Stephanie Bauer (stellvertretende Vorsitzende),
Prof. Dr. Ir. Pieter Jansens, Prof. Dr. Laurens Kuipers

Das Forschungszentrum Jülich stellt zurzeit auf einen neuen Zertifikatsanbieter zum digitalen Signieren von E-Mails um. Während dieser Umstellungsarbeiten kann es vorkommen, dass das DFN Community PKI Zertifikat, mit dem diese E-Mail signiert worden ist, als ungültig angezeigt wird.