[Beowulf] commercial clusters

[Jim Lux]

At 08:32 AM 9/29/2006, Vincent Diepeveen wrote:


>The above is the biggest problem. That's why you need good software that
>'hides' the supercomputer. Basically they want with their windows PC start
>your program (see attachment) and click somewhere and then run on a big
>supercomputer (for the average guy on the street a big supercomputer is
>everything that 2 hands cannot lift).
>
>If it was possible to build your own cluster in easy manner and then run for
>example a chessprogram at it in a user friendly way,
>there would be 100k+ clusters right now of 64 cpu's and more.
>
>Especially now that 1 chip hardly gets faster, we must think in 
>future more and more
>in clustered manner.
>
>Just doing simple math what i own myself:
>
>1996 : 200Mhz p6
>2006 : 2.4Ghz opteron
>
>both are 3 instructions per cycle processors.
>So hardware guys won exactly factor 12 in raw processing speed.
>
>Moore's law: each 18 month doubling in transistors.
>deduction from that: doubling in speed.
>
>Would mean we are faster now a 106 times  ( 2 ^ 6.66 )
>So somewhere a factor 8 is missing.
>
>The average user feels very well that there is no doubling in speed 
>each 18 months.
>
>So offering solutions to normal users to get their favourite program 
>executed faster in an user friendly manner is definitely a big market.

And this is what companies like Orion are targeting.  Specific 
classes of users who need lots of computation, with a specific 
program, and with $10K to $100K to spend.  I don't know that there 
are "consumer" applications like this yet, but in the generalized 
engineering world, there are a number of finite element codes for 
electromagnetics, structures, CFD, and thermal analysis that are 
widely used.  These codes take quite a lot of computation, so 
something that is a truly "turnkey" speed up by a factor of 10 or 100 
is worth it.  Turns a "load and run the model overnight" kind of 
operation into a "run the model while I get coffee" operation.


Jim