Re: Stop During PsN
Dear Kajsa,
Thanks very much for your kind help. However there's something I am not
very clear about the second point you mentioned above and I hope that you
could explain a little bit more.
Previously when I wanted to resume a PsN run I used the -directory= in the
command, and according to the PsN project homepage, this would enable a
resume of a PsN run after stop. Such a method works fine without the
-clean=3 option, but initiated a new search for covariates if I have
-clean=3 option. However, this method seems to be different from the above
mentioned.
It seems to me that the method you mentioned requires 2 steps. The first
one is to have only the included relations at the time of stop in the scm
file. Should I use the -directory= option in the command line?
Secondly, if I start a new run in a new directory, why should I use the
exactly same original input? By doing so, would be program being able to
skip the already established covariates and start from there onwards? What
if I use the unidentified relationships in the new run, and using a mod
file with relations in the scm logfile at the time of stop?
One last small question is that a few days ago, an error message saying
"psn.lst" is not created [something like this] was put forward by the
program during the middle of searching. The cov-par relation was completed
successfully before crush. Do you have any idea what might be the possible
reasons of this?
I am sorry for throw so many questions, but would highly appreciate it if
you could help me out of this. Many thanks.
Best Regards
Xinting
Quoted reply history
On 19 May 2014 15:01, Kajsa Harling <[email protected]> wrote:
> Dear Xinting,
>
> 1. This seems like an error during the NONMEM run itself. I suggest you
> run the same control stream with parallel nmfe72/73 on its own, without
> PsN, to see if the problem persists. If the model runs fine in parallel
> without PsN then please contact me directly and I will help you diagnose
> the problem. If the errors occurs independently of PsN then I hope someone
> else can help you.
>
> 2. You do not need to rerun the 13 iterations. To resume an interrupted
> scm it is always best, regardless of clean option, to set
> [included_relations] in the scm config file to the set of relations that
> were included at the time of the interruption. To see which those relations
> are you look in the scm logfile in the top level of the scm directory. Then
> you start a new scm run in a new directory with the new scm config file and
> the *same* input control stream as in the original run (no relations added
> in the input control stream). For details and syntax of the
> [included_relations] section please see the scm userguide.
>
> Best regards,
> Kajsa
>
>
>
> On 05/17/2014 05:03 PM, Xinting Wang wrote:
>
> Dear all,
>
> I would very much appreciate it if you could help me solve the below
> problem.
>
> 1. During the running of PsN, I came across a problem where no
> computation takes place. After the usual conduction, during some step the
> process seemed to be stopped. Although for parallel processing the worker1
> to worker5 directory was created, nothing followed Monitoring of Search in
> the OUTPUT, psn.lst file. This condition continued for 24 hours. I am not
> sure if this is a problem coming from the code or some errors with the
> computation system. Has anybody else come across a similar issue?
>
> 2. To reduce the hard disk burden I used -clean=3 in the scm process.
> However, after the process was stopped, it was impossible to resume the
> program using the statement -directory. Is it still possible resume the
> previous task? If so, how to? Repeating 13 rounds of problem is certainly
> something nobody want to do.
>
> Thanks a lot.
>
> --
> Xinting
>
>
> --
> -----------------------------------------------------------------
> Kajsa Harling, PhD
> System Developer
> Department of Pharmaceutical Biosciences
> Uppsala University
> [email protected]+46-(0)18-471 4308
> http://www.farmbio.uu.se/research/researchgroups/pharmacometrics/
> -----------------------------------------------------------------
>
>
--
Xinting