[Termtools] jambox?

[Jörg Endrullis]

Last year I have optimized Jambox for 60 seconds on a fast machine
(>= Athlon 64 3200+). With more than 60 seconds I did not do any
experiments, but felt free to use an "improved" strategy a) giving more
time to every single method and b) trying SAT matrix interpretations on
large self-labeled systems. The problem being b), it consumes much too
much time, then Jambox runs into timeout and the loop checker,
dependency pairs and matchbounds are not even started. Thus ignoring the
timeout would have been a better strategy, but I thought it would be 60
anyway :)

Joerg

Johannes Waldmann wrote:
> Peter Schneider-Kamp wrote (I guess it was meant to go to the list)
>> It seems to be 120 seconds of usertime. The version of AProVE that
>> we submitted is optimized for 60 seconds of wall time on a single
>> core machine.
> 
> 
> That's what I did, but I simply ignore the timeout parameter
> (because it was said that it would be 60 anyway)
> 
> It seems jambox does take it into account,
> and draws wrong conclusions.
> (It seems it thinks it has more time than it really gets?)
> 
> Johannes.
> 
> 

_______________________________________________
Termtools mailing list
Termtools at lists.lri.fr
http://lists.lri.fr/mailman/listinfo/termtools