Re: [Swarm-Modelling] True concurrency?

[Marcus G. Daniels]

Valmir,

> Is Swarm concurrent (agents run concurrently) or implements threaded 
> schedules l
With the increasingly prevalence of multicore CPUs and hyperthreading, I 
think it is starting to make more sense to implement parallel scheduling 
in an ABM toolkit.  For example, Niagara systems from Sun will have 32 
threads in flight.  There are now systems with 8 dual core Opterons 
available as well.   These systems could show linear or superlinear 
speed-ups of agent-based models.  Otherwise, Beowulf style clusters have 
at best latency on the order of a few microseconds, which is a long time 
to wait for an agent on another computer to perform a small action like 
a single agent's `step' method.