large clusters and topologies

[Steffen Persvold]

Patrick GEOFFRAY wrote:
> "highly scalable" again :-)
> 

[snip]

> In reality, the biggest SCI cluster that i know is in Paderborn :
> 96 nodes. I have talked in Mannheim with people who worked on it,
> and it seems to be working as 3 clusters of 32 nodes.

Not true, they have one 32 node system and one 96 node system. I was
personally there to set it up, so I ran a Bisection bandwidth test on
all 96 nodes with MPI (numbers are from mpptest from the MPICH suite):

Bisection bandwidth

In the bisection test the complete system is logically divided into two
subsystems and the aggregated bandwidth between the two subsystems is
measured. An example of this is splitting the system in two horisontally
and letting each  node in the upper half communicate with a node in the
lower part on a one-to-one basis. The numbers are produced using
mpptest. 

The total aggregated application network bandwidth for 96 nodes was
measured to:
Bandwidth:  3124.29 MByte/s at size 524288  bytes(peak)

They also recently won a GrandMaster Chess Competition with their chess
program P.ConNerS see http://www.scali.com/news/news_000721.html 

BTW: Aerospatiale also has a system (4 nodes)

Best Regards,
-- 
  Steffen Persvold               Systems Engineer
  Email : mailto:sp at scali.no     Scali AS (http://www.scali.com)
  Tlf   : (+47) 22 62 89 50      Olaf Helsets vei 6
  Fax   : (+47) 22 62 89 51      N-0621 Oslo, Norway