Statistics – Computation
Scientific paper
Dec 2010
adsabs.harvard.edu/cgi-bin/nph-data_query?bibcode=2010agufmsm51a1756o&link_type=abstract
American Geophysical Union, Fall Meeting 2010, abstract #SM51A-1756
Statistics
Computation
[0560] Computational Geophysics / Numerical Solutions, [2753] Magnetospheric Physics / Numerical Modeling, [7833] Space Plasma Physics / Mathematical And Numerical Techniques
Scientific paper
Global MHD/Hall-MHD magnetospheric models are not able to capture the full diversity of scales and processes that control the Earth's magnetosphere. In order to significantly improve the predictive capabilities of global space weather models, new CPU-efficient algorithms are needed, which could properly account for ion kinetic effects in a large computational domain over long simulation times. To achieve this much expected breakthrough in hybrid (particle ions and fluid electrons) simulations we developed a novel asynchronous time integration technique known as Discrete-Event Simulation (DES). DES replaces conventional time stepping with event processing, which allows to update macro-particles and grid-based fields on their own timescales. This unique capability of DES removes the traditional CFL constraint on the global timestep and enables natural (event-driven) coupling of multi-physics components in a global application model. We report first-ever parallel 2D hybrid DES (HYPERS) runs and compare them with similar time-stepped simulations. We also discuss our undergoing efforts on developing efficient load-balancing strategies for future 3D HYPERS runs on petascale architectures.
Catalyurek U. V.
Karimabadi Homa
Omelchenko Yu. A.
Saule Erik
No associations
LandOfFree
Parallel Event-Driven Global Magnetospheric Hybrid Simulations does not yet have a rating. At this time, there are no reviews or comments for this scientific paper.
If you have personal experience with Parallel Event-Driven Global Magnetospheric Hybrid Simulations, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Parallel Event-Driven Global Magnetospheric Hybrid Simulations will most certainly appreciate the feedback.
Profile ID: LFWR-SCP-O-1473286