Christian Schulte wrote: > Then, of course: runtime and memory. A word about memory: If you are just benchmarking different models in Gecode/J, the peak memory figures should be good enough for relative comparisons. They do not contain the overhead caused by the Java interface, however, so they are useless when comparing to other systems. And you should probably also present them as relative numbers.
Guido _______________________________________________ Gecode users mailing list [EMAIL PROTECTED] https://www.gecode.org/mailman/listinfo/gecode-users
