swarm-support
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: Instantiating multiple ModelSwarms


From: Marcus G. Daniels
Subject: Re: Instantiating multiple ModelSwarms
Date: 09 Aug 2001 15:48:53 -0600
User-agent: Gnus/5.070084 (Pterodactyl Gnus v0.84) Emacs/20.7

>>>>> "FW" == Fred Wan <address@hidden> writes:

FW> I can't find the information on what happens to the command-line
FW> parameters anymore, especially not of the ones used for swarm
FW> initialization itself. -b sets the swarmGUIMode parameter to 1,
FW> but are there any internal changes to swarm itself?

-b avoids initialization of the graphical support code. 
That's  significant because it means that on a Unix machines the simulation
won't attempt to open sockets to the X server (the display).  

For example, if you didn't use -b (and even if you also didn't have an
ObserverSwarm) you'd be unable to run your simulation on a remote
machine on a private network (say a node on a Beowulf cluster) because
on startup a connection to the display would be attempted, which would fail.

FW> Yep, that's right, but ordinary Swarms can be nested, so why not GUISwarms?

No one has demonstrated the will to make it so.

For me, I use GUI things mostly as a way to convince myself that
things are working as intended and to get a sense of boundary
conditions and parameter sensitivities.  Once I have this comfort
level and a game plan for sweeps, the sweeps themselves are all done
in parallel on multiple machines in batch mode (usually via a simple program
that spawns via rsh, respawning as machines in the set return their results).

                  ==================================
   Swarm-Support is for discussion of the technical details of the day
   to day usage of Swarm.  For list administration needs (esp.
   [un]subscribing), please send a message to <address@hidden>
   with "help" in the body of the message.



reply via email to

[Prev in Thread] Current Thread [Next in Thread]