[Beowulf] Cluster Applications in Submarine Environment

[Mark Hahn]

> > > contact me directly.  I would particularly like to hear from anyone doing
> > > parallel DSP, acoustic signal processing, or cryptography.  I'd also be
> > > interested in comments regarding low power per cubic foot packaging since I
...

> 1) For a given computation, nonprogrammable hardware (e.g. the ALU in a DSP 
> processor or even in a microcomputer) will be faster and/or lower powered 
> than programmable hardware (e.g. FPGA).  The tradeoff is that the 
> generalpurpose CPU widget might have stuff you don't need (and that you 
> could leave out of the FPGA).  Do not believe, without seriously checking, 
> claims of "you can put N PowerPC cores on a model Y FPGA" and think that 
> this means its the same as N PowerPCs.. Lots of little details like memory 
> and peripheral management, caches, lookahead pipelines, etc.

but the poster mentioned three specific workloads which are eminently
and commonly implemented tidily using FPGAs.  not things that would 
sanely choose to implement with N PPC cores on an FPGA.

> 2) FPGAs are by no means easy to program (if you want to achieve the 
> potential performance/power advantages). The tool sets are more expensive 

have you looked at the newer generation of environments?  I was impressed
by the Mitrionics stuff at SC05.  I have no idea whether their result was 
"efficient enough" to justify FPGAs, but it was certainly easy to do.
they appeared to have a fast development cycle, and were indeed based on 
xilinx with many-MB/s loader interfaces.  admittedly, I was just grazing
on the demoware, but it seemed to hold together.

my real point was that the Orion approach (cool-but-slow cpus connected
with a lightweight network) is really ideal for embarassingly parallel
applications, and development or dynamic environments.  it's a great tool,
but just because you've got a hammer, it's not all nails ;)