Alexis Arnaudon
Alexis Arnaudon
I see, but I'm not using MPI for this, I do many single cell simulations, using simple multiprocessing and each neuron run in a different isolated process. But never mind,...
In fact, having neuron connecting to X11 could be 'bad'. I was trying to run many small neuron computations on multiple nodes with dask (or even starting independent multiprocessing jobs...
Just in case, I noticed that the warning message I get many times is here: https://github.com/neuronsimulator/nrn/blob/6d23a393d8836ae852b39dc69d156b7a2d1d0f58/src/ivoc/ivocmain.cpp#L661
Unfortunately it's not quite possible, as we run many evaluations, and only a very few fail. It is not just one parameter that is critical, but a particular combination of...
``` export NEURON_MODULE_OPTIONS='-nogui' ```
Hello! You can use this: https://neurom.readthedocs.io/en/stable/morph_stats.html which has a configuration dictionary to select which features you want. The doc is for a cli, but you can use the python function...
@AurelienJaquier , just give me general feedback on that, I'll add tests/docs as for the other one later
@AurelienJaquier , good to review!
@AurelienJaquier I found the issue here, it was that the implementation of AHP_depth_slow was ignoring the first and last APs, I'm not sure why at all. Just playing with indices...
mmh I'm not sure this is the reason, as this feature start detecting APs after stim_start, no? it should catch the same AP as AHP_depth IMO, if we set the...