[Scip] Determinism

[Tobias Achterberg]

SCIP should be deterministic, provided that in each of the two runs you use
- the same computer and operating system,
- the same parameter settings,
- the same binary,
- the same problem instance (i.e., the exact same input file).

Of course, if you have set a time limit and SCIP hits this time limit, then the result is
also non-deterministic.

If you still have non-determinism, then there is a bug.


Tobias

Alejandro Burzyn wrote:
> Hello everyone, I'm currently using the SCIP solver and I've found
> recently that different runs
> on the same lp give me different times (with great difference, like 1h
> to 2h). At first I thought
> it was my fault because I was using other programs while runnning SCIP.
> But, when I read the
> stats report, there are differences not only in the times values,
> they're everywhere.
> So my question is: is SCIP deterministic? if it is not, is it possible
> to make it deterministic?
> Thank you very much for your attention,
> Alejandro Burzyn
>  
> 
> 
> ------------------------------------------------------------------------
> 
> ¡Obtén la mejor experiencia en la web!
> Descarga gratis el nuevo Internet Explorer 8
> http://downloads.yahoo.com/ieak8/?l=e1
> 
> 
> ------------------------------------------------------------------------
> 
> _______________________________________________
> Scip mailing list
> Scip at zib.de
> http://listserv.zib.de/mailman/listinfo/scip