building a grid

building a grid

Postby zeta » Thu May 10, 2007 6:52 pm

As part of my research (and also as a side project ;) ) I am designing and building a functioning grid. A bare bones grid distributed across the web may be constructed around the globus toolkit:

http://www.globus.org

and the flavor of MPI to suit:

http://www3.niu.edu/mpi/



issues:

a)security

b)interoperability (windows/linux/mac etc)

c)bandwidth



I'll post here some of the progress and specific problems I've come across
User avatar
zeta
 
Posts: 9
Joined: Sat Feb 17, 2007 10:49 pm

Postby Cuchulainn » Mon May 14, 2007 10:56 am

Zeta,

We now have some Monte Carlo code running on OpenMP and soon on MPI.



We would like to extend to a grid; what are the steps?
User avatar
Cuchulainn
 
Posts: 676
Joined: Mon Dec 18, 2006 2:48 pm
Location: Amsterdam, the Netherlands

Postby Cuchulainn » Mon May 14, 2007 10:58 am

With a duo-core setup we see _speedup_ of 1.8 (80%).



At some stage shared models will reach a limit and then we should use a distributed model (possibly in conjunction with sharing). is this correct?
User avatar
Cuchulainn
 
Posts: 676
Joined: Mon Dec 18, 2006 2:48 pm
Location: Amsterdam, the Netherlands

Postby zeta » Tue Mar 24, 2009 2:52 pm

woops, forgot about this thread :) 'tis probably a little late, but if you like we could run something here -> http://ligo.aset.psu.edu/pleiades/index.shtml



MC would be very easy to do..
User avatar
zeta
 
Posts: 9
Joined: Sat Feb 17, 2007 10:49 pm


Return to MPI/OpenMP

Who is online

Users browsing this forum: No registered users and 1 guest

cron