An international research team used a lattice-based method to identify a hard-to-detect pulsar in NASA data, showing how a different search strategy can reduce the computing time needed to find periodic signals.
Analyzing 15 years of observations from NASA’s Fermi Large Area Telescope, the researchers identified PSR J1622−7204, an isolated pulsar rotating 261 times per second. The study was published Nov. 19 in Physical Review Letters.
The new search took about 10 CPU hours. The paper contrasts that with a previous semicoherent search that failed to detect the pulsar after using about 600 CPU years. The paper’s abstract does not state a million-fold speedup; the precise comparison depends on how the computational costs are defined.
The algorithm maps the signal search onto a high-dimensional lattice, where researchers look for a short vector. This gives the search a mathematical route to identifying a pulsar’s periodic signal.
Gamma-ray searches face a very low photon count: typically, about one photon is detected for every 100 million pulsar rotations, the paper says. The authors say their approach may help researchers find more gamma-ray pulsars, including those in binary systems, where orbital motion complicates searches without detailed information about the system’s orbit.
The study’s authors are affiliated with the Max Planck Institute for Gravitational Physics, Leibniz Universität Hannover, Weizmann Institute of Science and Chalmers University of Technology. Supplementary material includes an example implementation, data and a notebook for reproducing the pulsar’s timing solution.



