new SGI Origin & Onyx 3k?

[Eugene Leitl]

W Bauske writes:
 > My point in all this is that I know the sorts of nodes Greg
 > uses and they are state of the art. So, when he says he's
 > using 1000+ nodes, he's describing on the order of 1.5 TFLOP of peak
 > performance. How many people would routinely solve single
 > problems using a TFLOP class cluster? Using 128 state of the
 > art processors, you're looking at 192 GFLOPS. How many people
 > routinely use 192 GFLOPS or more to solve a single problem at 
 > a time?

Actually, this begs for a questionnaire. I think the list is large and
diverse enough that we can ask actual users as to 1) which generation
of CPUs they are using 2) the amount of nodes they're using 3) whether
single or multiuser jobs 4) what is their problem domain 5) how much
memory 6) which interconnect technology 7) please add here anything
you deem interesting (but please do not let this snowball into Spanish
Inquisition).

If anyone is willing to create a webform for this questionnaire/poll
(egroups.com probably doesn't allow nonmembers to vote), we can also
ask CCL (computational chemistry list), which almost exclusively
consists of users.

Anybody is willing to give it a spin? I think the data would come very
handy for cluster vendors on-list, too.