www.neuron.yale.edu

Hello,

I'm running a set of NEURON simulations on a i686 linux system. At the beginning of my program, it opens a new file and then uses .aopen to append data to it as the simulation progresses. When I run my program without the actual simulations (so it finished running pretty much instantly) the I/O seems to work fine, and it saves the default data in the output file properly. However, when I run the program with the actual simulation I get an error saying that the file wasn't opened. The longest time I've been able to run it without getting the error is about an hour. Also, I can run the simulation to completion without saving the data. Therefore, I don't think there's a bug in my program since it can run the simulation to completion, and can also I/O within a short period of time (i.e. in case of running without the actual simulation). I'm a new unix user, and it might as well be a problem of the system. I was wondering if anyone else also faced such a problem and if you had a solution for it.

Yamin

The default hypothesis is that all "bugs" are user errors until proven otherwise.

noorubm wrote:then uses .aopen to append data to it as the simulation progresses . . .
I don't think there's a bug in my program since it can run the simulation to completion, and can also I/O within a short period of time

That's not the test, though, is it? The critical test is to try your file writing code on a toy problem, i.e. open a file, write data to it, close the file, then use aopen to reopen it for appending more data, close again, etc.