LAM/MPI General User's Mailing List Archives

From: Roberto Pasianot (pasianot_at_[hidden])
Date: 2003-04-09 11:12:53

• Next message: choy hau yan: "LAM: hi"
• Previous message: Tobias Wenzel: "Re: LAM: Haphazard Results with LAM-MPI"
• In reply to: Tobias Wenzel: "Re: LAM: Haphazard Results with LAM-MPI"
• Next in thread: Jeff Squyres: "Re: LAM: Haphazard Results with LAM-MPI"

 Hi, just my $0.02 worth

On Wed, 9 Apr 2003, Tobias Wenzel wrote:

> [ snipped ]
> Try to increase the calculation complexity (e.g. a biger input matrix).
> If the problem is the one I described this should make the times a bit
> more equal to each other.
>

 Also check your new cluster first e.g. with "netpipes" tool,
 simple and straight. This will tell you the behavior from the
 communications point of view.

 Cheers,

         Roberto

• Next message: choy hau yan: "LAM: hi"
• Previous message: Tobias Wenzel: "Re: LAM: Haphazard Results with LAM-MPI"
• In reply to: Tobias Wenzel: "Re: LAM: Haphazard Results with LAM-MPI"
• Next in thread: Jeff Squyres: "Re: LAM: Haphazard Results with LAM-MPI"