www.neuron.yale.edu

Hi,

If I am creating 10 cells with round robin algorithm used with parallel context, how do the cells distribute? To be precise, will the cells distribute like Cell[0], Cell[1], Cell[2], Cell[3], Cell[4] on one process and again Cell[0], Cell[1], Cell[2], Cell[3], Cell[4] on another process. Iteration of i works like 0, 2, 4, 6, 8 and 1, 3, 5, 7, 9 but not the cell distribution.

But my question is with reference to cell instantiation. If I give Cell=new refcell(), duplicate entries are created on each process as explained earlier. Kindly please explain how to avoid duplicate entries? Thank you.

Duplication is avoided by assigning each cell its own global identifier and using this global identifier to ensure that each processor gets only those cells that belong on it. For examples please read
Hines, M.L. and Carnevale, N.T.
Translating network models to parallel hardware in NEURON.
J. Neurosci. Methods 169:425-455, 2008.
and examine the source code for the models described therein.
Preprint is available from http://www.neuron.yale.edu/neuron/nrnpubs
Full source code is available from ModelDB https://senselab.med.yale.edu/modeldb/ via accession number 96444