Contributions à l'amélioration de l'extensibilité de simulations parallèles de plasmas turbulents

Fabien Rozar 1, 2
Abstract : Energy needs around the world still increase despite the resources needed to produce fossil energy drain off year after year. An alternative way to produce energy is by nuclear fusion through magnetic confinement. Mastering this reaction is a challenge and represents an active field of the current research. In order to improve our understanding of the phenomena which occur during a fusion reaction, experiment and simulation are both put to use. The performed experiments, thanks to Tokamaks, allow some experimental reading. The process of experimental measurements is of great complexity and requires the use of the most advanced available technologies. Currently, these measurements do not give access to all scales of time and space of physical phenomenon. Numerical simulation permits the exploration of these scales which are still unreachable through experiment. An extreme computing power is mandatory to perform realistic simulations. The use of High Performance Computing (HPC) is necessary to access simulation of realistic cases. This requirement means the use of large computers, also known as supercomputers. The works realized through this thesis focuses on the optimization of the Gysela code which simulates a plasma turbulence. Optimization of a scientific application concerns mainly one of the three following points : (i ) the simulation of larger meshes, (ii ) the reduction of computing time and (iii ) the enhancement of the computation accuracy. The first part of this manuscript presents the contributions relative to simulation of larger mesh. Alike many simulation codes, getting more realistic simulations is often analogous to refine the meshes. The finer the mesh the larger the memory consumption. Moreover, during these last few years, the supercomputers had trend to provide less and less memory per computer core. For these reasons, we have developed a library, the libMTM (Modeling and Tracing Memory), dedicated to study precisely the memory consumption of parallel softwares. The libMTM tools allowed us to reduce the memory consumption of Gysela and to study its scalability. As far as we know, there is no other tool which provides equivalent features which allow the memoryscalability study. The second part of the manuscript presents the works relative to the optimization of the computation time and the improvement of accuracy of the gyroaverage operator. This operator represents a corner stone of the gyrokinetic model which is used by the Gysela application. The improvement of accuracy emanates from a change in the computing method : a scheme based on a 2D Hermite interpolation substitutes the Padé approximation. Although the new version of the gyroaverage operator is more accurate, it is also more expensive in computation time than the former one. In order to keep the simulation in reasonable time, diferent optimizations have been performed on the new computing method to get it competitive. Finally, we have developed a MPI parallelized version of the new gyroaverage operator. The good scalability of this new gyroaverage computer will allow, eventually, a reduction of MPI communication costs which are penalizing in Gysela.
Document type :
Theses
Complete list of metadatas

https://tel.archives-ouvertes.fr/tel-01271032
Contributor : Abes Star <>
Submitted on : Monday, February 8, 2016 - 5:02:55 PM
Last modification on : Thursday, January 11, 2018 - 6:22:35 AM
Long-term archiving on: Saturday, November 12, 2016 - 2:20:12 PM

File

ROZAR_FABIEN_2015.pdf
Version validated by the jury (STAR)

Identifiers

  • HAL Id : tel-01271032, version 1

Citation

Fabien Rozar. Contributions à l'amélioration de l'extensibilité de simulations parallèles de plasmas turbulents. Modélisation et simulation. Université de Bordeaux, 2015. Français. ⟨NNT : 2015BORD0211⟩. ⟨tel-01271032⟩

Share

Metrics

Record views

364

Files downloads

259