It is not just HPC applications that are demanding low latency networks.
what applications are you thinking of? the only I can think of would be
lock/metadata traffic for very large parallel clustered DB/filesystem,
and even there the argument is weak.
I'm thinking of applications like on-line video games (first-person shooters,
real-time strategy, massively multi-player role-playing, etc. benefit from
lower latencies) and automatic electronic stock trading to name a few.
humans are deadly slow, and tend to be separated by geograpic latency.
the first hop on my cable connection is is 7 ms (perhaps not excellent),
and all latency aspects of human perception are in that range. I have
a hard time seeing how games would notice the difference between 60
and 30 us (readily achievable by Gb), let alone even detect 1-4 us latency
of current HPC interconnect...
unless you're suggesting that games use collectives akin to MPI
broadcast/reduce/etc?
I'm guessing that online auto-trading has "think time" that dominates
1-100 us network latency. again, geography and network topology would
probably make more difference than shaving 30 us down to 10...
regards, mark hahn.
_______________________________________________
Beowulf mailing list, Beowulf@beowulf.org
To change your subscription (digest mode or unsubscribe) visit
http://www.beowulf.org/mailman/listinfo/beowulf