[Comp-neuro] Special topics at the NEURON course

Ted Carnevale carnevalet at sbcglobal.net
Sun Oct 5 17:46:26 CEST 2008


A few seats are still available for this year's NEURON course
at the annual SFN meeting, which will take place on Friday,
Nov. 14, in Washington, DC.  For the course description and
online registration form, see http://www.neuron.yale.edu/dc2008.html
The registration deadline is Friday, Oct. 17--less than two weeks
from today--so you should act quickly to sign up.

As in prior years, the course will cover a broad span of topics
that range from how to get started, to expert level tips for the
most productive usage of NEURON.

In addition, this year's course will present the latest advances
in two very rapidly moving areas:

1.  Using Python with NEURON.  Until recently, models have been
specified with hoc and/or NEURON's GUI tools.  But now it is
possible to use Python in conjunction with hoc, or even instead
of hoc.  This is very important because it puts a very powerful,
modern programming language in the hands of NEURON users, and
gives them access to the large, rapidly expanding, and freely
available libraries of scientific software written in Python.

2.  Parallel simulations.  Over the past few years, NEURON gained
the ability to take advantage of parallel computer hardware in order
to accelerate simulations.    Efficient parallel simulation, with
speedup proportional to the number of processors, is now possible
for a wide range of problems on a wide range of parallel hardware.
The most recent advance, with perhaps the greatest potential for
benefiting the largest number of users, is multithreaded execution
of simulations on multicore workstations.  For some time now, all
new Macs, and most new PCs, come with multicore processors.  Attend
the NEURON course to find out how to make your simulations take
advantage of all that latent processing power!


More information about the Comp-neuro mailing list