Aurora capable of computing 2 billion billion calculations per second
LEMONT, Ill. (WLS) -- This is what scientists at Argonne National Laboratory in Lemont call a node: six huge graphics processors and two large CPUs cooled with water to make major calculations a cinch.
Argonne's new supercomputer doesn't just have one node, 10 or 100, instead it has 10,000 of them. Each single rack of nodes weighs eight tons and are cooled by thousands of gallons of water.
Its computing power equals 2 exaflops, or 2 billion-billion calculations per second.
"That is, if every person on the planet was doing math 24/7, 365 days a year for just a little over a year, is what Aurora can do in one second," explained Michael Papka, the Argonne Leadership Computing Facility Director.
Aurora is the massive new supercomputer that's housed by Argonne, just 30 minutes away from the city. Costing hundreds of millions of dollars, it aims to accelerate advancements across an array of scientific research, creating complex climate models, cracking cosmic conundrums and contributing to cures for chronic conditions. In other words, it has the potential to change the course of American science.
"It will do traditional things like climate studies. It will do things like the evolution of the universe, and then we will bring it to the more practical side. Things like designing better airplane engines, better airfoils, better drug design. It's really meant to support a wide variety of things," Papka said.
It's not even fully operational yet but it's already the second fastest computer in the world. It takes up the floor space of four tennis courts and it's right in our backyard.
When it's fully turned on later this year, Aurora is expected to be the fastest computer on Earth, armed with artificial intelligence to help scientists make informed decisions faster.
"I can try many, many different 'what if' scenarios for how the climate will play out," said Papka. "I could do all that in the computer that informs our decision and then of course we can choose where to go based on that. So it accelerates our ability to to think."
That includes climate scientists like Rao Kotamarthi, who will use Aurora to train artificial intelligence weather models to help predict future climate issues on an extremely local level.
"Our goal is to get a really, really granular forecast with models like this on a global scale up to the next 50, to 100 years," said Kotamarthi.
Kotamarthi said he and other climate scientists like him will be able to use Aurora's sophisticated processing power to turn petabytes of data into transformative climate breakthroughs running any number of "what if" scenarios.
"Then, I can do lots of case studies like I can change this, I can change that and see how mitigation, adaptation and all these questions can be answered. So that's our holy grail," said Kotamarthi.
Even potentially giving neighborhood specific modeling to predict the likelihood of flooding or other extreme weather events so cities and companies can manage mitigations needed now.
"While we are reducing carbon dioxide, there are still climate changes happening. How do I adapt to this? How do I make my infrastructure resilient? These are the questions we asking," said Kotamarthi.