A one-hour media briefing on Climateprediction.net
25 August 2004
ATTENTION: TECHNOLOGY, SCIENCE EDITORS
A one-hour media briefing on Climateprediction.net, the newest internet computing project to move to the University of California, Berkeley's new BOINC platform, which allows computer users to participate in many different public computing projects at once.
Climateprediction.net, which runs models of global climate change, joins SETI@home and Predictor@home, which predicts how proteins fold, on the BOINC (Berkeley Open Infrastructure for Network Computing) platform.
2-3 p.m. Thursday, Aug. 26
First-floor conference room, Space Sciences Laboratory addition, just off Grizzly Peak Blvd.
David Anderson, director of SETI@home and developer of BOINC
David Frame, project coordinator, Climateprediction.net
Tolu Aina, research officer, Climateprediction.net
Until now, people who wanted to donate their idle computer time to the search for intelligent life in the universe or the hunt for new drugs had to choose one or the other. BOINC lets users participate in as many distributed computing projects as they want. Climateprediction.net, launched a year ago and now boasting 62,000 participants in 130 countries, is the newest project to switch to BOINC, but it soon will be joined by other projects around the world, with some two dozen expected to switch to BOINC in the next year.
"With BOINC, you don't have to choose; you can participate in many different public prediction projects and control the amount of time each gets," said Anderson. "It uses computer time more efficiently, in case one of the projects is down, and allows users more than one pet project."
Climateprediction.net, which is run by several universities in the United Kingdom, requires substantial computing time because it repeatedly runs climate models to determine how they respond to tweaks in initial approximations. This helps scientists understand and improve the models so they more accurately predict the effects of changes in atmospheric carbon dioxide or sulfur.
The longest running distributed computing project is UC Berkeley's SETI@home, which is now transitioning its 5 million users to BOINC. The project Predictor@home, based at the Scripps Research Institute, harnesses the power of internet computing to predict a protein's 3-D structure from its amino acid sequence, a difficult problem whose solution could greatly impact biomedical research.