are super-computers still used today..

[Dennis Toy]

Back in the late-80's and early 90's, Cray Inc. had the faster supercomputer on earth, the Cray-2. Then Blue Gene came along in 1998 and beat a chess player in Russia. I asked this because PC today have as much power as the Cray-2 and I read that super-computers are custom built. Are supercomputer still used by universitys, science-labs and designers.

[Admiral Valdemar]

Yes.

[Kazuaki Shimazaki]

Think of it this way, for a particular technological generation, whatever a personal computer can do, a supercomputer can beat due to sheer bulk. Thus, until we can solve the Grand Unified Theory on a laptop in a few seconds, and find the biggest prime numbers while we are at it, yes we would always have supercomputers.

[tharkûn]

Define supercomputer.

At work I have an SGI which kicks the utter crap out of your high end PC's, it also costs more than a car.

Our lab utilizes a distributed network which makes that look like utter crap for the truly huge datasets.

A couple places still build single box supercomputers, but pricing is making it ridiciously more cost effective to distribute the workload over a network. You tend to be able to afford more computing power, you can slowly add capacity, and you have a longer operational lifetime.

Today's supercomputer tends to be hundreds of smaller computers working in concert and there is no end in sight for how big scientists want to make that network.

Think of it this way, for a particular technological generation, whatever a personal computer can do, a supercomputer can beat due to sheer bulk. Thus, until we can solve the Grand Unified Theory on a laptop in a few seconds, and find the biggest prime numbers while we are at it, yes we would always have supercomputers.

Barring massive advances in quantum computing, and possibly even then, such a computer would have to violate the laws of physics. Modeling macroscopic phenomena from the level at which GUT would be applicable would require well in excess of 10^23 bits for nontrivial cases. Using current prime hunting techniques, even a converting Mars into one giant computer will eventually reach a limit at prime detection.

[The Grim Squeaker]

Nowadays Grid computing (an infinite amount of computers each carrying out a piece of the calculations) is gaining more and more ground thanks to stuff like google’s toolbar.

[Admiral Valdemar]

Also, big business use mainframes today which are better at holding data and processing it than straight up supercomputers that typically look at one problem and compute it. There are degrees of power in this area, but it's safe to say that computers vastly more powerful than any single one in a person's home do exist.

[Dennis Toy]

Define supercomputer.

At work I have an SGI which kicks the utter crap out of your high end PC's, it also costs more than a car.

You know your right, a super-computer is a relative term, because todays super-computer could become tommorows home pc.

As i said in the beginning, your High-End PC has as much computing power as a Cray-2 made in 1991.

[Admiral Valdemar]

By the way, don't you mean Deep Blue in your first posts about Kasparov's defeat? Blue Gene is new and used only for mapping protein folding, as far as I can tell.

[tharkûn]

You know your right, a super-computer is a relative term, because todays super-computer could become tommorows home pc.

Supercomputer is defined by the price tag

Your PC, given current trends, is years behind my SGI. A good part of that is because the SGI is optimized to crunch numbers - GUI is optional, machine code is RIS, and I don't even want to think about the memory handling; but even without that the CPU is still massive (literally - anything larger and cooling becomes an issue). The architecture in the SGI is complete overkill for the home market, it will never be used in the home PC.

It isn't just about computing power, a good number cruncher is for more than a PC writ large. Eventually the home PC will have dense enough circuitry to beat my SGI, but it will never go through the QC, optimization, and egineering.

[Dennis Toy]

yea deep blue

[Dennis Toy]

by the way,...whats an SGI

[TrailerParkJawa]

by the way,...whats an SGI

Silicon Graphics is a company located in Mountain View California. They are known for producing very powerful workstations for graphics work. However, they are a sick little company now. Don't buy their stock, its been in the tank for years and the company has no real leadership. Of course this is probably more than you wanted to know, but hey, what the heck.

[Dennis Toy]

the same SGI who made the computers that made the terminator, jurassic park, forest gump and Star Wars the Phantom Menace, now in the tank? We gotta bail it out....

[WyrdNyrd]

the same SGI who made the computers that made the terminator, jurassic park, forest gump and Star Wars the Phantom Menace, now in the tank? We gotta bail it out....

Why? In a capitalist world, you adapt to your environment or you die.

I must confess, I have often wanted to see struggling companies succeed, because of some sentimental reason. MG Rover would be one example. And it is always sad when people lose their jobs because their firm went under.

But at the same time, sentimental support can encourage laziness and arrogance in company management (as if the typical exec needed any encouragement to be lazy and arrogant.)

Progress can only come by constant improvement and adaptation, not by resting on your laurels.

[Zero]

Progress can only come by constant improvement and adaptation, not by resting on your laurels.

Actually, much of what we call progress is people simply trying to apply their laziness in a practical way by simplifying the work, making their lives easier.

[The Grim Squeaker]

You mean design or faster boot ups?