poepping at cmu.edu
Thu Jan 29 22:57:01 EST 2004
>From what I know of the VT "Cluster", there really wouldn't be any value of
using 'cluster software' for argus. The code itself is quite single-threaded,
but since it's stream/filter-based, you can easily already set up a pipeline
for complicated operations. To my knowledge, there isn't any "Apple Cluster"
software, people use the various GRiD utilities to manage jobs on the
multi-computer made up of a bunch of G5's (i.e. it's not like a VAX-cluster,
or the other "cluster" technologies)..
> -----Original Message-----
> From: owner-argus-info at lists.andrew.cmu.edu [mailto:owner-argus-
> info at lists.andrew.cmu.edu] On Behalf Of eric-list-argus at catastrophe.net
> Sent: Thursday, January 29, 2004 1:36 AM
> To: Carter Bullard
> Cc: argus-info at lists.andrew.cmu.edu
> Subject: Re: MacOS Experiences
> On Wed, 2004-01-28 at 12:37:46 -0500, Carter Bullard proclaimed...
> > I'm getting better performance using G5's than Xeon's
> > primarily due to the little endian issues on Intel machines.
> > Based on the traffic mix, as much as 3x performance.
> > I've never seen a Mac under perform a comparable Intel
> > box as an Argus probe, and I have seen it out perform
> > on a regular basis.
> Excellent -- that helps alot.
> I know I asked before about clustering, but have you heard of anyone
> running argus on a cluster of xserv's or the g5 servers? I'm curious
> to know if it would work (I haven't seen much on the clustering
> capabilities yet, but I'm doing my research).
> Thanks again.
> - eric
More information about the argus