Dear Lionel,
Then indeed the numpy-array-based approach by Alberto is the best bet at the moment. You probably will need to do this block-wise to not overload memory. Through an EBRAINS infrastructure voucher, we currently develop a NEST backend for the SONATA format. So if your specification is stored in SONATA format, you could use that once we have a prototype ready.
Otherwise, we could into the possiblities of the conn_gen interface and libneurosim to add a backend that could read your files directly.
Best, Hans Ekkehard
--
Prof. Dr. Hans Ekkehard Plesser Head, Department of Data Science
Faculty of Science and Technology Norwegian University of Life Sciences PO Box 5003, 1432 Aas, Norway
Phone +47 6723 1560 Email hans.ekkehard.plesser@nmbu.no Home http://arken.nmbu.no/~plesser
On 20/10/2021, 10:26, "kusch lionel" lionel.kusch@univ-amu.fr wrote:
Dear Hans,
The "specific function" is a complex function written in Matlab using morphology of the neurons for defining the connectivity between neurons. The output of this function is a file with pre-synapsy and prost-synapsy GIDs. For the moment, there is not plan to integrate this algorithm in NEST.
My question is in the case where the connection of synapse is defined in a file (GID of input and GIDs of output), what is the best way to load it in NEST and connect the neurons?
Best regard,
Kusch Lionel
On 19/10/2021 22:16, Hans Ekkehard Plesser wrote: > Dear Lionel, > > I assume by "specific function" you mean a function that receives lists of potential source and target neurons and then decides which neuron to connect to which in a mathematical or algorithmic way? > > Could you share more information about how your connectivity is defined? If you do not want to share that information on this public mailing list, feel free to send a support request with more details to support@ebrains.eu and mention that it should be routed to me. > > Best, > Hans Ekkehard > > -- > > Prof. Dr. Hans Ekkehard Plesser > Head, Department of Data Science > > Faculty of Science and Technology > Norwegian University of Life Sciences > PO Box 5003, 1432 Aas, Norway > > Phone +47 6723 1560 > Email hans.ekkehard.plesser@nmbu.no > Home http://arken.nmbu.no/~plesser > > > > > On 18/10/2021, 12:09, "kusch lionel" lionel.kusch@univ-amu.fr wrote: > > Dear all, > > In my working group, we are working on the creation of the spiking > neural network of human CA1 (4.8 million neurons). We are using a > specific function for defining the connection between the neurons. This > function gives us 30 billion pairwise connections. > We would like to know what the most efficient for the creating the > network in NEST 3 using this pairwise connectivity. > > Thank you in advance for your answer, > > Kusch Lionel > > _______________________________________________ > NEST Users mailing list -- users@nest-simulator.org > To unsubscribe send an email to users-leave@nest-simulator.org > > _______________________________________________ > NEST Users mailing list -- users@nest-simulator.org > To unsubscribe send an email to users-leave@nest-simulator.org