Loosely coherent searches for sets of well-modeled signals

We introduce a high-performance implementation of a loosely coherent statistic sensitive to signals spanning a finite-dimensional manifold in parameter space. Results from full-scale simulations on Gaussian noise are discussed, as well as implications for future searches for continuous gravitational waves and related searches for slowly modulated signals, such as emitted by radio pulsars. We demonstrate an improvement of more than an order of magnitude in analysis speed over previously available algorithms. As searches for continuous gravitational waves are computationally limited, the large speedup results in gain in sensitivity.