-
Physics informed Neural Networks applied to the description of wave-particle resonance in kinetic simulations of fusion plasmas
Authors:
Jai Kumar,
David Zarzoso,
Virginie Grandgirard,
Jan Ebert,
Stefan Kesselheim
Abstract:
The Vlasov-Poisson system is employed in its reduced form version (1D1V) as a test bed for the applicability of Physics Informed Neural Network (PINN) to the wave-particle resonance. Two examples are explored: the Landau damping and the bump-on-tail instability. PINN is first tested as a compression method for the solution of the Vlasov-Poisson system and compared to the standard neural networks.…
▽ More
The Vlasov-Poisson system is employed in its reduced form version (1D1V) as a test bed for the applicability of Physics Informed Neural Network (PINN) to the wave-particle resonance. Two examples are explored: the Landau damping and the bump-on-tail instability. PINN is first tested as a compression method for the solution of the Vlasov-Poisson system and compared to the standard neural networks. Second, the application of PINN to solving the Vlasov-Poisson system is also presented with the special emphasis on the integral part, which motivates the implementation of a PINN variant, called Integrable PINN (I-PINN), based on the automatic-differentiation to solve the partial differential equation and on the automatic-integration to solve the integral equation.
△ Less
Submitted 23 August, 2023;
originally announced August 2023.
-
Optilization of the gyroaverage operator based on hermite interpolation
Authors:
F Rozar,
C Steiner,
G Latu,
M Mehrenberger,
V Grandgirard,
Julien Bigot,
T Cartier-Michaud,
Jean Roman
Abstract:
Gyrokinetic modeling is appropriate for describing Tokamak plasma turbulence, and the gyroaverage operator is a cornerstone of this approach. In a gyrokinetic code, the gyroaveraging scheme needs to be accurate enough to avoid spoiling the data but also requires a low computation cost because it is applied often on the main unknown, the 5D guiding-center distribution function, and on the 3D electr…
▽ More
Gyrokinetic modeling is appropriate for describing Tokamak plasma turbulence, and the gyroaverage operator is a cornerstone of this approach. In a gyrokinetic code, the gyroaveraging scheme needs to be accurate enough to avoid spoiling the data but also requires a low computation cost because it is applied often on the main unknown, the 5D guiding-center distribution function, and on the 3D electric potentials. In the present paper, we improve a gyroaverage scheme based on Hermite interpolation used in the Gysela code. This initial implementation represents a too large fraction of the total execution time. The gyroaverage operator has been reformulated and is now expressed as a matrix-vector product and a cache-friendly algorithm has been setup. Different techniques have been investigated to quicken the computations by more than a factor two. Description of the algorithms is given, together with an analysis of the achieved performance.
△ Less
Submitted 9 February, 2016;
originally announced February 2016.
-
Evaluating kernels on Xeon Phi to accelerate Gysela application
Authors:
G. Latu,
M. Haefele,
J. Bigot,
V. Grandgirard,
T. Cartier-Michaud,
F. Rozar
Abstract:
This work describes the challenges presented by porting parts ofthe Gysela code to the Intel Xeon Phi coprocessor, as well as techniques used for optimization, vectorization and tuning that can be applied to other applications. We evaluate the performance of somegeneric micro-benchmark on Phi versus Intel Sandy Bridge. Several interpolation kernels useful for the Gysela application are analyzed an…
▽ More
This work describes the challenges presented by porting parts ofthe Gysela code to the Intel Xeon Phi coprocessor, as well as techniques used for optimization, vectorization and tuning that can be applied to other applications. We evaluate the performance of somegeneric micro-benchmark on Phi versus Intel Sandy Bridge. Several interpolation kernels useful for the Gysela application are analyzed and the performance are shown. Some memory-bound and compute-bound kernels are accelerated by a factor 2 on the Phi device compared to Sandy architecture. Nevertheless, it is hard, if not impossible, to reach a large fraction of the peek performance on the Phi device,especially for real-life applications as Gysela. A collateral benefit of this optimization and tuning work is that the execution time of Gysela (using 4D advections) has decreased on a standard architecture such as Intel Sandy Bridge.
△ Less
Submitted 3 August, 2015; v1 submitted 16 March, 2015;
originally announced March 2015.
-
Non regression testing for the JOREK code
Authors:
Guillaume Latu,
Marina Becoulet,
Guilhem Dif-Pradalier,
Virginie Grandgirard,
Matthias Hoelzl,
G. Huysmans,
Xavier Lacoste,
Eric Nardon,
Francois Orain,
Chantal Passeron,
Pierre Ramet,
Ahmed Ratnani
Abstract:
Non Regression Testing (NRT) aims to check if software modifications result in undesired behaviour. Suppose the behaviour of the application previously known, this kind of test makes it possible to identify an eventual regression, a bug. Improving and tuning a parallel code can be a time-consuming and difficult task, especially whenever people from different scientific fields interact closely. The…
▽ More
Non Regression Testing (NRT) aims to check if software modifications result in undesired behaviour. Suppose the behaviour of the application previously known, this kind of test makes it possible to identify an eventual regression, a bug. Improving and tuning a parallel code can be a time-consuming and difficult task, especially whenever people from different scientific fields interact closely. The JOREK code aims at investing Magnetohydrodynamic (MHD) instabilities in a Tokamak plasma. This paper describes the NRT procedure that has been tuned for this simulation code. Automation of the NRT is one keypoint to keeping the code healthy in a source code repository.
△ Less
Submitted 16 November, 2012;
originally announced November 2012.
