• About the author

    Leigh Smith is a Senior Research Associate at the Institute of Astronomy, University of Cambridge, and a researcher at the Cambridge Astronomical Survey Unit. His work spans wide-field time-domain and astrometric surveys, exoplanet transit detection, stellar occultations and Galactic structure. He has a particular interest in low-mass stars and brown dwarfs, especially thick disk, halo and benchmark examples, and in completing the census of the solar neighbourhood.

    He works mainly on ESA’s PLATO mission. To cope with the volume of data PLATO will produce, he developed CETRA, a CUDA transit detection algorithm that uses a realistic transit model and separates a linear search in time from a periodic search over trial periods. At high cadence this makes it far faster than existing CPU codes, and it is now incorporated into the PLATO Exoplanet Analysis System.

    Alongside CETRA, he leads the VIRAC proper motion catalogues of the Galactic bulge and plane from VVV and VVVX near-infrared imaging, and is developing a deep image-stacking pipeline for measuring proper motions in crowded bulge fields, where GPU convolution is a natural fit. He also works on transit light curve modelling, including ring systems around exoplanets.

    • Publications

      CETRA: A fast, sensitive exoplanet transit detection algorithm implemented for GPUs, Smith, L. C., Ahmed, S., De Angeli, F., et al. 2025, Monthly Notices of the Royal Astronomical Society, 539, 297–306. arXiv:2503.20875 [https://arxiv.org/abs/2503.20875]

    • Acknowledgements

      We would like to thank Leigh Smith for his contributions and for sharing his expertise with our course participants.

      This deep dive is part of the DiRAC GPU Performance Portability Project which has been funded through the UKRI Digital Research Infrastructure Programme.

  • Overview

  • 1. Finding Earth-like planets

  • 2. GPU programming foundations

  • 3. The CETRA algorithm

  • 4. Performance and results