Beowulf & FFT

[Walter B. Ligon III]

--------

> For mpich 4 processors are slower than 2 processors. It seems that
> communication between processes on the same node is so much faster that
> np=2 turns to be faster than np=4. However, the effect is the same in
> mpich-1.1.2 and mpich-1.2.0; the latter does shared-memory communication.
> mpipro doesn't show the same effect.
> 
> Since I'm in the process of expanding the beowulf, I'm wondering whether
> switching to 133 MHz would improve the results (given that I can find
> a motherboard that supports ECC - we had that discussion).
> 
> Any thoughts and comments on this are appreciated.

Generally, if the problem is that node-to-node communication is too slow,
upping your system bus speed isn't going to do much in the way of improving
your parallel system performance.

The first thing that strikes me is that a 400x400 operation is pretty small
for a Beowulf.  Its going to be hard to get good efficiency on a small
problem like that.

I am not familliar with the FFT library you have, but it may be that it was
not designed for a machine like a Beowulf, but rather for a finer grain
machine like some of the MPPs.  Thus the best approach may be a different
bit of software, or reworking that software.  All-to-All is probably NOT
the best implementation on an MPI running over TCP/IP.

Of course, you could always upgrade to a faster network if you have the $$$.

Walt

-- 
Dr. Walter B. Ligon III
Associate Professor
ECE Department
Clemson University