soma voltage becomes "nan" during simulation

[eacheon]

Hello,

My vector recording membrane voltage gets "nan" values after some point process (synapses) is triggered. Other than the error that the recorded vector has values "nan" from a certain index on, there's no any other error or warning messages.

After nrniv finished its runs:

Code: Select all

oc> .... (no warning or error message)
oc> soma.v
         nan

I got a screenshot of the voltage window, but here I find no way to attach it in messages.

Still trying to debug into this. It was fine the day before. Just wondering any one have met such a case and would like to share some insight?

- e

[ted]

It was fine the day before.

So something changed between then and now. Unless somebody else has access to
your machine, you changed it. If you have been keeping track of revisions of your code,
it should be fairly easy to go back to the last good version of your program and fix the
problem.

[eacheon]

It was fine the day before.

So something changed between then and now. Unless somebody else has access to
your machine, you changed it. If you have been keeping track of revisions of your code,
it should be fairly easy to go back to the last good version of your program and fix the
problem.

exactly, exactly.

I made some mistake in the initialization of the model. It really helps to read thoroghly throught that part in reference again. Also as you said, an inspection of the backups indicated where the errors are introduced.