Originally posted by: HWFN_Hossein_Hojabri

Hi,

I'm embedding CP in an algorithm to solve subproblems iteratively. The algorithm is working fine, however, for large number of iterations (100,000 for example), it has memory problems since it could take even around 37 GB (in CP) in a few instances. I don't think that it's normal forCP to take this much memory.  I think there must be a way to release the memory without having to reallocate the env, cp and the model for every single iteration. Could you help me decrease the memory without sacrificing much the CP runtime?

A little about my algo: I declare the env in the main function, and pass it as a pointer to CP (*env) to solve the subprolems. I use cp.getHeap() a few time, and IlcRevInt a few times. Please let me know if I should provide more info.

Thanks!

Originally posted by: rdumeur

Hi,

Did you try using valgrind or purify to check for memory leaks?

Otherwise, we need a more precise description of the problem. It would be best if you could isolate a pathological case where memory growth can be observed.

Normally, each time endSearch() is called, all memory allocated during search should be released.

Also, could you please confirm the version of  CP (aka Solver/Scheduler/Dispatcher) you are using?

Thank you in advance,

Cheers,

Originally posted by: HWFN_Hossein_Hojabri

Hi,

Sorry for my delay. The problem is solved and it is working perfectly.

Thanks,