On 2012-05-10, Sara Pashmin <sarapashm at googlemail.com> wrote:
> I run a maxima program parallel on two computers.
> The program consists of 3 loops, which the data (with a lot of
> if-conditions and equations) have to been run 76x101x101 times (I hope
> that I explain it in English clearly).
>
> The simulation on one of PCs was finished yesterday after 12 hours. On
> the other one the simulation has been run since 3 days but it seems to
> me that the simulation has been stopped but not broken.
Sara, you'll have to provide some additional information in order for
others to understand the problem. What is the output of build_info() on
each computer? What is the program which you are running?
Without understanding anything, my first suggestion is to print the loop
index and any other relevant data in the outer loop (the 76 x loop), so
that you can see if you're making progress.
Second suggestion is to break up the whole computation so that you don't
have to run the whole thing in one pass. For example: load the current
state for the outer loop, run the inner loops, save the current state,
then exit Maxima. Then if you have to kill the program, you haven't lost
much, and you can review the saved states.
Those are just some ideas that can be helpful when making long
computations.
best
Robert Dodier