[gecode-users] Peak memory and recomputation

[Grégoire Dooms]

Christian Schulte wrote:
> Dear Grégoire,
>
> Mikael, Guido, and me have been looking into your example: it seems to be
> okay. The reason is due to last alternative optimization. Just by shere luck
> with c_d = 4, the search engine is able to just keep a single space on the
> stack.
>
>   
> So what you have been observing is no bug but a feature. That said, also for
> a problem that small and the search tree that shallow the memory figures are
> close to meaningless.
>
>   
Ok.
I still don't understand why the peak memory is higher when fewer 
solutions are searched (with the same other parameters). It might be an 
implementation detail of the search engine. I'll have a look at it.
Thanks for your help.
Best,
--
Grégoire