[Beowulf] help on building Beowulf

Mark Hahn hahn at MCMASTER.CA
Tue Nov 20 13:15:57 PST 2007


> According to my experiences to run HPC applications in Shanghai Super
> Computing Center. Myrinet interconnection brought to many failure with even
> a small application. All users are crazy with the interconnections and we
> had to restart the applications once and once again.

I mean no slight of SSC users or admins, but to me this sounds like 
some sort of mistake.  temperatures, bad cables, etc.  our largest 
myrinet cluster is 287 nodes, versus ~512 for SSC's cluster, but we 
do have a greater number of total ports.  I guess it's conceivable 
that SSC's problem is specific to their very large switch...



More information about the Beowulf mailing list