ARECIBO: The Einstein@Home project recently discovered a pulsar 15 kilometers in diameter located more than 30,000 light years from Earth by analyzing radio wave data from the Arecibo Observatory in Puerto Rico. Image: COURTESY OF CORNELL UNIVERSITY
A massive distributed computing network known as the Einstein@Home project has made its second big celestial find in the past six months—a pulsar 15 kilometers in diameter located more than 30,000 light years from Earth. Not a bad track record for a network that runs on processing power donated from computers worldwide. The finding represents one of many discoveries that Einstein@Home's creator expects in the coming years.
Einstein@Home is a project that allows computer users (whether individuals or organizations) to contribute their PCs' unused computing power to the greater cause of helping scientists search for pulsars, which are spinning neutron stars (aka collapsed, super-dense stars) that emit pulses of radio waves, x-rays or gamma rays. Last week alone, about 100,000 PCs contacted Einstein@Home servers located at the Max Planck Institute for Gravitational Physics in Hannover, Germany, the University of Wisconsin–Milwaukee and elsewhere to download a few megabytes of data at a time for analysis or upload analyzed data.
The project was set up in February 2005 as way of searching for gravitational waves (also emitted by pulsars) in data collected by the Laser Interferometer Gravitational-Wave Observatory (LIGO) detector instruments located in Livingston, La., and on the Hanford Nuclear Reservation near Richland, Wash. Gravitational waves, which have never been detected directly, were predicted by Albert Einstein in 1916 on the basis of his theory of general relativity, although he thought they were too weak to be observed, says Einstein@Home Director Bruce Allen, a physicist at Max Planck and U.W.–Milwaukee.
"Pulsars are the cores of old stars that are crushed by gravitational forces," Allen says. "They get crushed into a ball that basically—if you could crush this ball to be a factor of three smaller, it would collapse into a black hole. It's the last stage of matter before you collapse into a black hole." Such objects allow researchers to study the behavior of matter at extremely high pressures and densities.
By tapping into computers signed up for the project, researchers are given access to more processing power than was available through their own computers or any other computer system that the National Science Foundation could afford to pay for, Allen says. With gravitational wave data, "it's a situation where the sensitivity of your search is determined by how much computing power you have," he adds. "More computing power lets us search more deeply in the LIGO data."
Given the uncertainty of finding gravitational waves, Allen and his team in 2009 expanded the Einstein@Home program to search for binary pulsars by analyzing radio-wave data from the Arecibo Observatory in Puerto Rico. Through this work, Einstein@Home researchers located two pulsars: The first, found in August, is a disrupted binary pulsar 17,000 light-years away rotating more than 40 times a second that had broken free of its binary companion. The most recent pulsar, identified as PSR J1952+2630, is surprisingly similar in some ways to the earlier pulsar, Allen says. "The second pulsar is located in roughly the same part of the sky, and it's spinning at roughly the same rate," he says. Yet the new pulsar is more unusual than the first because it's still part of a binary star system, but its companion star is a white dwarf star that is more massive than is typically the case. Allen and his colleagues published their findings Monday on arXiv.org.
Whereas the Arecibo data gives Einstein@Home researchers and volunteers a more attainable goal, Allen is hoping that upgrades to the LIGO instruments combined with faster PC processing power will enable the project to detect gravitational waves within the next decade. Nearly half of these volunteer PCs already have superfast graphics processing units (GPUs) in addition to central processing units (CPUs). GPUs, originally designed for rendering highly sophisticated video game graphics, can now be programmed to process information between 10 and 20 times faster than CPUs. "It's like we've got about another a half million functioning machines," on the network, Allen says.
Working with leading GPU maker NVIDIA Corp., Allen and his team have improved Einstein@Home software so that it can use GPUs more efficiently. "The detection of gravitational waves from new neutron stars using LIGO data will be a combination of increasingly sensitive LIGO instruments and an increase in the use of GPUs," Allen says.