A forum for Spin users

You are not logged in.

#1 2011-06-22 07:56:41

Registered: 2011-06-22
Posts: 1

Spin Memory Usage Issue


I have a question about the memory usage of spin.  I am attempting verify some basic LTL properties on models with 100+ variables.  At the moment, all non-determinism in the model is during initialization.  After initialization each path is deterministic in nature (the rest of the model is enclosed in a d_step) and fairly short (depth < 20).  However, running the models I am running out of memory (5 GB is available to Spin).  The only explanation I can see is if Spin is storing every reachable state in the model.

My understanding of spin is that once a path is fully explored and fails to provide a counter-example, the path is discarded.  This should lead to memory usage linear to path depth.  Is this the behavior I should be observing?  If so, is there a reason I am not seeing it?

Thank you,


#2 2011-06-22 15:06:54

Registered: 2010-11-18
Posts: 691

Re: Spin Memory Usage Issue

states are never discarded, because any state at all could be revisited -- and it would significantly worsen the search complexity of the algorithm if we discarded states. (from linear it would become quadratic).
there are lots of ways to do searches while using less memory though: try -DCOLLAPSE, or -DHC4, or -DBITSTATE for example


Board footer

Powered by FluxBB