mance with thousands of messages
per second -- this will always be your bottleneck.
As such, I believe it's unfair (and imprecise) to use DNS performance results
for a group communication (MPI) environment comparison. I had good
experiences with MPI + FreeBSD, although don't have
the IQ to understand why do you keep using FreeBSD if it makes
you unhappy, doesn't support the hardware you bought/have, perform poorly on
most situations you have to deal with, and you think its developers don't
have a clue of what they're doing.
--
Ricardo Nabinger Sanchez <[EM