www.neuron.yale.edu

Hi all,

When running our model on Macs with a powerpc chip, it runs fine with variable dt.

When running on i686's (either Mac or Linux) we get the following error, and it will only run with variable dt off. Any ideas would be greatly appreciated.


CVode-- At t = 0 and h = 1.65133e-10, the corrector
convergence failed repeatedly or with |h| = hmin.

CVode 26287b4 acell_home_ advance_tn failed, err=-7.
err=-7
/Applications/NEURON/nrn-5.9/i686/bin/nrniv: variable step integrator error
near line 10
{run()}
^
fadvance()
advance()
step()
continuerun(2000)
and others

Please send me a zip file with all the hoc,ses,mod files required to reproduce the problem.
michael dot hines at yale dot edu

Dr. Hines mailed me the following:

"The problem is with

access acell_home_
L=0

in create_mspcells.hoc

Set L to a nonzero value or else get rid of acell_home_ completely
(in 5.9 and above it is no longer necessary to place ARTIFICIAL_CELL
instances in such a dummy section)"

Removing the acell_home_ instance solved the problem without affecting the simulation.