Accelerated Searches of Gravitational Waves Using Graphics Processing Units

Abstract

Abstract. The existence of gravitational waves was predicted by Albert Einstein. Black hole and neutron star binary systems will product strong gravitational waves through their inspiral and eventual merger. The analysis of the gravitational wave data is computationally intensive, requiring matched filtering of terabytes of data with a bank of at least 3000 numerical templates that represent predicted waveforms. We need to complete the analysis in real-time (within the duration of the signal) in order to enable follow-up observations with some conventional optical or radio telescopes. We report a novel application of a graphics processing units (GPUs) for the purpose of accelerating the search pipelines for gravitational waves from coalescing binary systems of compact objects. A speed-up of 16 fold in total has been achieved with an NVIDIA GeForce 8800 Ultra GPU card compared with a standard central processing unit (CPU). We show that further improvements are possible and discuss the reduction in CPU number required for the detection of inspiral sources afforded by the use of GPUs.