It's amazing the lengths some companies will go to to produce the world's fastest supercomputer. But what's even more amazing is that these so-called supercomputers really aren't that super. Sure, they perform calculations in an extremely fast manner, but what's really happening here is you've got thousands of processors linked together to form one big computer. It's not a new concept, but the fact that these mini-computers are linked together to form a supercomputer is somewhat baffling. Don't get me wrong - I understand the idea behind supercomputers - but I fail to see how they are all that important.
In the article, it's mentioned that the NEC Earth Simulator is the fastest computer and it is currently in Japan. Apparently, the United States government, with nothing else to worry about, thinks that we're losing our high-tech prowess. Give me a break. Recently, at Virginia Tech, thousands of Power Mac G5's were strung together to create a supercomputer. All the government needs to do is buy a bunch of these computers, hook them up in parallel and then viola!, we've got the fastest supercomputer again, and this time, at a fraction of the price of some IBM product.
Perhaps the government could learn a little something from PBS - even though this article is almost two years old, it's still quite relevant and if I had the cash, I'd go about building my own mini-supercomputer and put it to work on d.net projects. Speaking of which, I wonder how long it'd take IBM's newest supercomputer to crack RC5-72. d.net has been working on it for almost a year and is just barely over 1/2% completion.
URL: IBM gives glimpse of Blue Gene performance | CNET News.com
In the article, it's mentioned that the NEC Earth Simulator is the fastest computer and it is currently in Japan. Apparently, the United States government, with nothing else to worry about, thinks that we're losing our high-tech prowess. Give me a break. Recently, at Virginia Tech, thousands of Power Mac G5's were strung together to create a supercomputer. All the government needs to do is buy a bunch of these computers, hook them up in parallel and then viola!, we've got the fastest supercomputer again, and this time, at a fraction of the price of some IBM product.
Perhaps the government could learn a little something from PBS - even though this article is almost two years old, it's still quite relevant and if I had the cash, I'd go about building my own mini-supercomputer and put it to work on d.net projects. Speaking of which, I wonder how long it'd take IBM's newest supercomputer to crack RC5-72. d.net has been working on it for almost a year and is just barely over 1/2% completion.
URL: IBM gives glimpse of Blue Gene performance | CNET News.com



0 Comments:
Post a Comment
<< Home