PIConGPU: Getting Particle-in-cell Plasma Simulations Ready for Exascale


PIConGPU: Getting Particle-in-cell Plasma Simulations Ready for Exascale

Debus, A.; Widera, R.; Lebedev, A.; Bastrakov, S.; Pausch, R.; Steiniger, K.; Pöschel, F.; Stephan, J.; Bussmann, M.

No official abstract here. We present an overview talk on laser-particle acceleration research with the challenges of modeling laser-plasma interactions using PIConGPU on latest exascale machines (JUWELS BOOSTER, Perlmutter,Summit and early-access systems for Frontier).We show the softwarestack of the particle-in-cell code PIConGPU (alpaka,openPMD, ...) with an emphasis on performance-portability and I/O at exascale. Furthermore we showed in an example how detailed performance benchmarking and profiling (together with JSC) helped improving code performance.

Keywords: particle-in-cell code; PIConGPU

  • Invited lecture (Conferences) (Online presentation)
    10 Year Anniversary Workshop of NVIDIA Application Lab at Jülich, 21.-22.6.2022, Jülich, Deutschland

Permalink: https://www.hzdr.de/publications/Publ-36295