A pulsar spinning 261 times every second was hiding in 15 years of NASA observations, and a new search method found it with about 10 hours of computing on a standard processor. The same hunt would have taken an estimated 600 years with the previous best approach.
Pulsars are rapidly spinning neutron stars that send beams of radiation through space. When a beam sweeps past Earth, it can be detected like the flash of a lighthouse. Finding them in gamma-ray data is especially hard: NASA’s Fermi Gamma-ray Space Telescope detects only a few photons from any one pulsar, amid signals from other sources and background noise. Researchers also have to work out the star’s spin, how that spin is slowing, and its location in the sky. Searching all the possible combinations can demand more computing power than is available.
An international team used a method inspired by cryptography to find PSR J1622−7204 in Fermi data. The technique was developed by study co-authors Dotan Gazith and Barak Zackay of the Weizmann Institute of Science in Israel. Rather than test every possible signal directly, the researchers turned the search into a geometry problem: finding a short vector in a high-dimensional lattice, a regularly spaced set of points. Related algorithms were developed by cryptographers studying ways to break encryption designed to resist future quantum computers. The team used them here to search for a star, not to crack a code.
Each run took about five minutes on one computer core. After the method identified a candidate, the researchers checked it against the full Fermi dataset. The signal was strong enough to rule out a chance detection. The results were published Thursday in Physical Review Letters.
The speedup could change how scientists search for pulsars. Until now, teams often divided Fermi observations into smaller pieces and relied on volunteers’ spare computing power. Even with that help, finding one pulsar could take the equivalent of hundreds of computer-years. The new technique makes searches possible on ordinary computers and could help researchers find objects that older methods missed.
The team’s next target is pulsars in binary systems, where a pulsar orbits another star. The companion’s gravity changes the pulsar’s signal, making these systems especially difficult to detect. Colin Clark, an astrophysicist at the Max Planck Institute for Gravitational Physics and a study co-author, said the data may contain tens or even hundreds of pulsars that researchers have not yet found. The team hopes to combine the new algorithm with volunteer computing through its Einstein@Home program to look for them.
Open access to NASA’s observations made the search possible. Fermi launched in 2008 and scans the entire sky every three hours, observing gamma rays, the highest-energy form of light. After its first year of calibration, its data became publicly available and has continued to grow. Researchers outside the Fermi mission could therefore bring a different tool to the same archive. As more observations join the public dataset, the faster search gives scientists a practical way to keep looking for signals that were once buried in the noise.
