Performance of simulation

[mob@cs.rpi.edu]

Hello,

I'm now trying to estimate the gain in cutting a big topology into several
parts. When I cut the topology into 2 parts ( the traffic parameters
for the whole topology are the same as that for the partial topology ),
the time needed for a partial topology is much less than half of time
needed for the whole topology. I can't understand what cause this result,
can anyone give me suggestions?

As I can guess, one factor may be the data location of the simulation, so
I need to check how many cache, memory and virtual memory pages are used
for each simulation. can anybody tell me how I can do this?

Thanks

Bin Mo