Recent EUROfusion Achievements in Support of Computationally Demanding Multiscale Fusion Physics Simulations and Integrated Modeling. (3rd October 2018)

• Record Type: Journal Article
• Authors: Voitsekhovitch, I.; Hatzky, R.; Coster, D.; Imbeaux, F.; McDonald, D. C.; Fehér, T. B.; Kang, K. S.; Leggate, H.; Martone, M.; Mochalskyy, S.; Sáez, X.; Ribeiro, T.; Tran, T.-M.; Gutierrez-Milla, A.; Aniel, T.; Figat, D.; Fleury, L.; Hoenen, O.; Hollocombe, J.; Kaljun, D.; Manduchi, G.; Owsiak, M.; Pais, V.; Palak, B.; Plociennik, M.; Signoret, J.; Vouland, C.; Yadykin, D.; Robin, F.; Iannone, F.; Bracco, G.; David, J.; Maslennikov, A.; Noé, J.; Rossi, E.; Kamendje, R.; Heuraux, S.; Hölzl, M.; Pinches, S. D.; Da Silva, F.; Tskhakaya, D.;
• Is Part Of: Fusion science and technology. Volume 74:Number 3(2018)
• Page Start: 186
• Page End: 197
• Publication Date: 2018-10-03
• Subjects: High-performance computer -- infrastructure for integrated modeling -- code optimization and parallelization; Fusion reactors -- Periodicals; Nuclear fusion -- Periodicals; Fusion reactors; Nuclear fusion; Periodicals; 621.48405
• DOI: 10.1080/15361055.2018.1424483
• Languages: English
• ISSNs: 1536-1055

Abstract

Abstract: Integrated modeling (IM) of present experiments and future tokamak reactors requires the provision of computational resources and numerical tools capable of simulating multiscale spatial phenomena as well as fast transient events and relatively slow plasma evolution within a reasonably short computational time. Recent progress in the implementation of the new computational resources for fusion applications in Europe based on modern supercomputer technologies (supercomputer MARCONI-FUSION), in the optimization and speedup of the EU fusion-related first-principle codes, and in the development of a basis for physics codes/modules integration into a centrally maintained suite of IM tools achieved within the EUROfusion Consortium is presented. Physics phenomena that can now be reasonably modelled in various areas (core turbulence and magnetic reconnection, edge and scrape-off layer physics, radio-frequency heating and current drive, magnetohydrodynamic model, reflectometry simulations) following successful code optimizations and parallelization are briefly described. Development activities in support to IM are summarized. They include support to (1) the local deployment of the IM infrastructure and access to experimental data at various host sites, (2) the management of releases for sophisticated IM workflows involving a large number of components, and (3) the performance optimization of complex IM workflows.