April 24, 2013
Researchers using supercomputer to improve satellite precipitation observation
UNIVERSITY PARK, Pa. , April 24 — A research team including Penn State engineers will use one of the world's most powerful computers to enhance scientists' understanding of global precipitation.
The Penn State-Princeton University-Aerospace Corp. team is one of only 28 teams granted access to the Blue Waters supercomputer, the National Science Foundation (NSF) recently announced.
The NSF-funded Blue Waters machine, located at the National Center for Supercomputing Applications at the University of Illinois, is capable at peak performance of nearly 12 quadrillion floating point operations per second and has demonstrated a sustained system performance of more than one petaflop on a range of commonly used science and engineering applications, according to the agency.
The team will receive 140 million core hours on the Blue Waters' 512,000 computer cores over the next year.
To put matters in perspective, most high-end laptop computers are equipped with only four processing cores.
"If you filled Beaver Stadium to capacity with everyone using the best laptop you could buy, Blue Waters would still be faster," said Patrick Reed, associate professor of civil engineering. "You can do 10 years of computing in 10 minutes with Blue Waters. We can do hundreds of years of computations in a day."
The research team will use Blue Waters to tackle the problem of large gaps in precipitation data for large parts of the world. The goal, he said, is to help scientists and researchers move toward an integrated global water cycle observatory, something that the World Climate Research Program has long been seeking.
"The U.S. and Europe have really good weather data," Reed said. "We take it for granted we have good weather forecasts. But that data costs more than $1 billion a year. There are countries in the world with few, if any, observations."
That lack of data on other parts of the planet makes it challenging to understand and predict droughts or floods.
"Water is a significant source of uncertainty and risk in global development," Reed said.
That's where Blue Waters comes in. "It's a bit of an oversimplification, but we're going to look at how well these satellites can cover the earth," the civil engineer stated.
The Blue Waters computations and simulations will take into account factors such as spending limits, international coordination limits, satellite lifespan and other tradeoffs when it comes to re-tasking satellites.
The researchers will use the existing Global Precipitation Measurement mission as their technical baseline.
"This will be one of the largest computational experiments in history in merging the astrodynamics design with global hydrology."
In the end, Reed said he hopes Blue Waters can be used to create a new satellite constellation for a more complete precipitation observation of the planet.
Reed has used supercomputing on a number of previous research efforts, including the Texas Advanced Computing Center's Ranger system to evaluate watershed models from around the country, including the Susquehanna River basin.
Source: Penn State
10/30/2013 | Cray, DDN, Mellanox, NetApp, ScaleMP, Supermicro, Xyratex | Creating data is easy… the challenge is getting it to the right place to make use of it. This paper discusses fresh solutions that can directly increase I/O efficiency, and the applications of these solutions to current, and new technology infrastructures.
10/01/2013 | IBM | A new trend is developing in the HPC space that is also affecting enterprise computing productivity with the arrival of “ultra-dense” hyper-scale servers.
Ken Claffey, SVP and General Manager at Xyratex, presents ClusterStor at the Vendor Showdown at ISC13 in Leipzig, Germany.
Join HPCwire Editor Nicole Hemsoth and Dr. David Bader from Georgia Tech as they take center stage on opening night at Atlanta's first Big Data Kick Off Week, filmed in front of a live audience. Nicole and David look at the evolution of HPC, today's big data challenges, discuss real world solutions, and reveal their predictions. Exactly what does the future holds for HPC?