A5002872357- Solar and Space Plasma Dynamics
- Astro and Planetary Science
- Geomagnetism and Paleomagnetism Studies
- Parallel Computing and Optimization Techniques
- Computational Fluid Dynamics and Aerodynamics
- Distributed and Parallel Computing Systems
- Meteorological Phenomena and Simulations
- Tropical and Extratropical Cyclones Research
- Advanced Data Storage Technologies
- Galaxies: Formation, Evolution, Phenomena
- Astrophysics and Star Formation Studies
- Computer Graphics and Visualization Techniques
Aalto University
2017-2022
Collaboration et al., (2021). The Pencil Code, a modular MPI code for partial differential equations and particles: multipurpose multiuser-maintained. Journal of Open Source Software, 6(58), 2807, https://doi.org/10.21105/joss.02807
Modern compute nodes in high-performance computing provide a tremendous level of parallelism and processing power. However, as arithmetic performance has been observed to increase at faster rate relative memory network bandwidths, optimizing data movement become critical for achieving strong scaling many communication-heavy applications. This gap further accentuated with the introduction graphics units, which can by multiple factors higher throughput data-parallel tasks than central units....
Abstract Magnetohydrodynamical (MHD) dynamos emerge in many different astrophysical situations where turbulence is present, but the interaction between large-scale (LSDs) and small-scale (SSDs) not fully understood. We performed a systematic study of turbulent driven by isotropic forcing isothermal MHD with magnetic Prandtl number unity, focusing on exponential growth stage. Both helical nonhelical was employed to separate effects LSD SSD periodic domain. Reynolds numbers ( <?CDATA...
Over the last ten years, graphics processors have become de facto accelerator for data-parallel tasks in various branches of high-performance computing, including machine learning and computational sciences. However, with recent introduction AMD-manufactured to world's fastest supercomputers, tuning strategies established previous hardware generations must be re-evaluated. In this study, we evaluate performance energy efficiency stencil computations on modern datacenter processors, propose a...
To understand solar and stellar dynamos combining local global numerical modelling with long-term observations is a challenging task: even state of the art computational methods resources, parameter regime remains unattainable. Our goal to relax some approximations, in order simulate more realistic systems, try connect results theoretical predictions state-of-the-art observations. We present here first test-field measurements from our higher-resolution runs improved heat conduction...