>Cray claimed that its T3E was capable of over one teraFLOPS. Needless
>to say, no one's ever managed to verify their claims (based, not
>incoincidentally, on aggregate performance).
I wont argue on this point.
>Neither NEC nor Fujitsu has published numbers which meet even the ASCI
>Red baseline. If they've recently released results to the contrary, I
ASCI Red vs NEC? Come on, its not fair to compare apples to oranges. NEC
isnt exactly backed by the DOE. That's a single example of whats possible
but there arent exactly commercially available 9000+ node intel cpu
supercomputers are there? (not yet atleast) If you look at the machines
Cray Research has compared to the machines NEC has, either statistically,
or by market share, NEC wins hands down. Why is it that every industrial
country in the world uses NEC supercomputers for weather prediction except
the United States? (this happens to be a field in desperate need of high
speed massivelly parallel machines as you know) Cray cant compete.
Its a bit late and I hate to argue, so Ill just leave it at this.
---------------------------------------------------------------------------
Send administrative requests to [EMAIL PROTECTED]
