pygwtf 1.0.0

Waveform generation and likelihood approximations for slowly-evolving GW sources in the time-frequency domain.

gwtf

Waveform generation and likelihood approximations for slowly-evolving GW sources in the time-frequency domain in O(10) microseconds.

Usage of waveform generator in an end-to-end parameter estimation provided in the notebook Parameter_estimation.ipynb

Currently being used to analyse the stellar-origin binaries in the LISA data-challenge Mojito lite.

Features:

• Support both CPU and GPU based computations of both waveforms and likelihoods.
• Supports inner-product/waveform batching over GPU threads.
• Contains search-specific implementation of semi-coherent detection statistic, designed to minimise memory use and maximise batching of sources (only valid for single mode waveforms).
• Response function that generalises to slowly varying arm-lengths.
• Response can use either analytic or supplied orbits (Mojito-orbit reading in functionality is present and functional but somewhat crude).

Waveforms Supported:

• TaylorT2Ecc (non-spinning)
• TaylorT3 (spinning)

Contributors

• Christian Chapman-Bird
• Diganta Bandopadhyay