On 8/7/05, Rajkumar S s_raj@flashmail.com wrote:
Apologies for digressing from the list charter...
Good that you found your way to the list. Now tell us what new hot projects are up and running at Sarovar ;)
Anand Babu wrote:
Actual design of supercomputer is very much dependent on the application needs. Best way to benchmark is to use the real world applications and measure performance.
One question that I always had about super computers are that, do they need specialized programs to take advantage of it's processing power? For example, I have a program for computing X that is running in my existing cluster (say Beowulf), can I run that application in thunder and take advantage of it's speed, or do I have to start writing it from scratch to run in thunder?
No where in super computer literature, I have read about writing applications for it, except for libraries like PVM. It will be interesting to know about how real world applications are written for supercomputers.
AFAIK, the Supercomputer that may be a cluster of 1000+ individual computers appears as a single system to its system administrator/end user. Chaos like applications manage your other applications. ( http://www.purehacking.com/chaos/ ) We could be using OpenOffice, Postgres or any other application as we would on any other system. This much I digressed from AB, and also, once I get my hands on such a machine, I'll update more details :)
Speaking about supercomputers replacing human brain, we really don't know enough about human brain. We don't even have a good acceptable definition for Artificial Intelligence. Its a long way to go to for computers to achieve human like intelligence (30-50 years may be).
As the quote goes, Artificial intelligence is no match for natural stupidity :)
Marvin Minsky pronounced AI to be "brain-dead" some time back, but that is not the same as being stupid :)