GPU acceleration of gravitational-wave signal models

– CI: Rory Smith

Abstract: This project will develop a GPU implementation of gravitational-wave signal models to alleviate the computational bottleneck in LIGOs parameter estimation codes. Parameter estimation is an optimal tool for gravitational-wave signal detection but its current high cost prohibits current codes to be used as a search pipeline. The dominant cost of the codes is in evaluating signal models which are compared to data to obtain the likelihood of the data containing a gravitational-wave. The signal models are highly parallelizable. Hence there is strong motivation to implement the models on a GPU. The cost of signal modelling would need to be made between one and two orders of magnitude cheaper to develop a cheap gravitational-wave search pipeline.

Searching for gravitational waves using parameter estimation tools would allow LIGO to search a much larger volume of the observable universe. Additionally, we expect to be able to detect the “stochastic background” of gravitational waves using only a day of data, compared to around 40 months of data with current techniques. Detection of this background would be a big astronomical prize, allowing the study of the whole population of binary black holes and neutron stars, as well as having cosmological implications.

The project will produce a GPU parallelized implementation of a commonly used gravitational- wave signal model known as “IMRPhenomP”. This implementation would allow us to benchmark expected performance improvements to current parameter estimation tools. Looking further, the implementation would serve as a corner stone of a fully GPU-implemented Monte Carlo algorithm, the core algorithm used in parameter estimation.

WordPress Theme built by Shufflehound.

Copyright © Astronomy Data and Compute Services

ADACS is delivered jointly by Swinburne University of Technology and Curtin University. ADACS is funded under Astronomy National Collaborative Research Infrastructure Strategy (NCRIS) Program via Astronomy Australia Ltd (AAL).