Recent technical developments in the EPW code

ORAL

Abstract

EPW is an open-source Fortran code to study electron-phonon interactions based on density-functional perturbation theory and maximally localized Wannier functions. Since its first release in 2016 as one of the core modules of the Quantum ESPRESSO suite, EPW has been widely used to investigate a variety of phonon-mediated quantum processes in real materials.

In this talk, we outline the recent technical developments in EPW. They cover both algorithmic improvements and highly scalable optimizations. To name a few, we fully exploited crystal symmetry to reduce the number of evaluations of electron-phonon matrix elements on coarse Brillouin-zone grids and we implemented multi-level MPI parallelizations combined with OpenMP to reduce communication overhead and improve strong scaling. Taken altogether, these and further additions constitute the first steps to enable the transition of EPW to exascale computations of the electron-phonon coupling.

*This work was supported by the U.S. Department of Energy (DOE), Office of Science, Basic Energy Sciences (BES) under Award DE-SC0020129 and computational resources were provided by the Oak Ridge Leadership Computing Facility at the Oak Ridge National Laboratory and the Texas Advanced Computing Center (TACC) at The University of Texas at Austin.

–

Presenters

  • Hyungjun Lee

    • Oden Institute for Computational Engineering and Sciences, University of Texas at Austin

Authors

  • Hyungjun Lee

    • Oden Institute for Computational Engineering and Sciences, University of Texas at Austin

  • Samuel Poncé

    • Theory and Simulation of Materials (THEOS), École Polytechnique Fédérale de Lausanne

  • Elena R Margine

    • Department of Physics, Applied Physics and Astronomy, Binghamton University-SUNY
    • Department of Physics, Applied Physics, and Astronomy, Binghamton University-SUNY
    • Department of Physics, Applied Physics and Astronomy, Binghamton University-SUNY – Binghamton, NY13902, USA
    • Department of Physics, Binghamton University-SUNY

  • Feliciano Giustino

    • Physics, University of Texas at Austin
    • University of Texas at Austin
    • Oden Institute for Computational Engineering and Sciences, University of Texas at Austin
    • Department of Physic, The University of Texas at Austin, Austin, Texas 78712, USA, Oden Institute for Computational Engineering and Sciences
    • Oden Institute, University of Texas at Austin
    • Department of Materials, University of Oxford
    • Department of Physics, University of Texas at Austin
    • ODEN Institute for Computational Engineering and Sciences, University of Texas at Austin