FNN vs GigabitEther & Myrinet

[Greg Lindahl]

On Wed, Oct 24, 2001 at 06:41:59AM -0400, Patrick Geoffray wrote:

> You are right. The GAMMA work is excellent, and the idea to provide 
> a OS-bypass protocol easily portable over the Ethernet family is a 
> good one.

Another project doing this is MVIA, which, coupled with MVICH, would
provide equivalent software functionality to GM plus MPICH/GM.

One big problem with both the GAMMA and MVIA projects is that
originally, neither was doing any kind of error correction, simply
telling the application "whoops! there's a problem." That doesn't work
with MPI programs. Now both have added error correction. But they both
still have the problem of the fast changing market in gigE chipsets,
and neither one has any scheme for directly plugging into BSD or Linux
kernel drivers.

It's kind of a shame, because for tiny clusters, the price is pretty
compelling: 8 port gigE switches are down to $950, and programs like
MM5 at typical resolutions get a big benefit from faster than fast
ethernet networks at 16 cpus.

greg