Hierarchical approach to matched filtering using a reduced basis

Searching for gravitational waves from compact binary coalescence (CBC) is performed by matched filtering the observed strain data gravitational-wave observatories against a discrete set of waveform templates designed to accurately approximate expected signal, and are chosen efficiently cover target search region. The computational cost scales with both number required parameter space in-band duration waveform. Both these factors increase in difficulty as current improve sensitivity, especially at low frequencies, may pose challenges third-generation observatories. Reducing would make searches future detector more tractable. In addition, it be easier conduct that incorporate effects eccentricity, precession or light sources (e.g. subsolar). We present hierarchical scheme based on reduced basis method decrease conducting matched-filter search. Compared methods, we estimate without any loss speedup factor ~ 10 signal-to-noise ratio (SNR) least =6.0, 6 SNR 5. Our dominated linear operations which highly parallelizable. Therefore, implement our algorithm using graphical processing units (GPUs) evaluate commercially motivated metrics demonstrate efficiency GPUs CBC searches. can extended generic allows efficient GPUs.