CCL: Wireless cluster

 Sent to CCL by: "Andrew D. Fant" []
 John McKelvey wrote:
 > Please don't laugh, or maybe this is so funny [ridiculous?] that you'll
 > get a good laugh, and it will make your day!
 > Anyway, I want to build a small linux cluster [6-8 total processors] but
 > don't have a lot of  cooling in one place.  Running wires would not be
 > practical, if not impossible  Now, the application is extremely
 > coarse-grain, and only a very, very small amount of data gets moved
 > about once initialization of a job is done..  Can it be done
 Hey John,
     I wouldn't make it my first solution, but it could be done.  As others have
 pointed out, your bandwidth would be limited, your latency would be rather
 variable, and you would be stuck in half-duplex mode for your communications.
 You might want to look at more sophisticated methods for data transmission to
 make up for this if the process isn't truly embarrassingly parallel.  In
 particular, coding with MPI even if it isn't strictly required for the
 parallelism might pay dividends, since you would be able to use gather/scatter
 and broadcast/reduce code to sling what data you have around, which might be
 more efficient than what you can do from the shell.  Alternately, if you break
 down and get a good wireless router that supports multicast well, you might want
 to look at using multicast to handle communications.  Again, the idea being to
 announce data that other systems might care about and let the ones  that
 actually DO care grab it from one set of packets.  Then again, they are doing
 wonders with pantone-matched drop cables and discrete baseboard mounting clips
 these days, too. ;-)
 Hope this helps,
 Andrew Fant    | And when the night is cloudy    | This space to let
 Molecular Geek | There is still a light          |---------------------- | That shines on me               | Disclaimer:  I don't
 Boston, MA     | Shine until tomorrow, Let it be | even speak for myself