SolveClimate is looking for writers, please click here to learn more.
iPod Supercomputer to Make Climate Science Energy-Efficient
by Teresa Herrmann - May 14th, 2008
in
Three researchers at the DOE Berkeley Lab have gotten techie attention because they found a way to build the world's most accurate and efficient climate prediction supercomputer by using the same microprocessor technology of the iPod.
But there's something more fascinating about why the research was needed in the first place.
The super machines used to predict climate change today are the Hummers of the computer world.
They're veritable electricity-guzzlers -- both from the megawatts they consume directly and for the gigantic cooling systems that are needed to deal with the mass heat they generate.
Still, the computers aren't powerful enough to handle rigorous climate observation.
According to the The Berkeley Lab researchers, the world needs a whole new class of supercomputers that are 1,000 times more potent than the ones used today.
The biggest problem with current machines involves cloud simulations -- the bread and butter of climate modeling. They can't model them in the detail that's called for.
So the researchers argue for a giant, one-kilometer cloud model as a fix. But that would cost $1 billion to build using today's machines and would require 200 megawatts of electricity to run -- enough juice to power a small city of 100,000 residents.
Enter iPod.
In the end, the researchers concluded that a supercomputer using about 20 million low-power "iPod" microprocessors would deliver the results and cost just $75 million to construct.
Best of all, the machine would consume less than four megawatts of power.
And the idea is moving ahead. The Berkeley Lab has already signed a collaboration agreement with Tensilica Inc. -- a global leader in processor technology -- to help design the new energy-efficient supercomputer, and it's working with scientists at Colorado State University to get a prototype system up and running.
At least in the world of supercomputers, the Hummer-to-Prius transition seems to be under way.
Source: Science Daily
Bookmark/Search this post with:
Bakken shale (sweet crude)
I have been following the Bakken Shale discovery in Montana, North Dakota, and Canada for quite some time. I was under the distinct impression that there was a huge oil discovery located at that site. More than the Saudi, and Iraq's largest fiields. Are the Economic Advisors wrong about the amount of sweet crude that is available? I read that there was nearly 400 Billion Barrels of oil available, is this all hype or is there any truth to this story? Now I have just read that only a months supply is all we may get from that field? It is difficult for me to believe that investors could be duped into a false assumption of so little oil available. Even nine months is better than nothing, however we sure need a great deal more fuel than just nine months supply. Is there more information available pertaining to the Bakken Crude? I truly hope that the writer of the Article that said quite the opposite of what I have been reading about the Bakken is totally false. I was under the impression that we really had hope for a little light at the end of the tunnel, and some relief from the high fee for gas at the pumps. Just what are facts about the Bakken? What if anything can we expect from this alledged huge find?
Also will the Solar Power be installed in Florida, a State know for it's Sunshine? If so, will it relieve some of the fossil fuel demands on the Citizens of that State, and aid in the reduction of fuel cost at the gas pumps?
Post new comment