Performance portability is considered to be an inevitable requirementin the exascale era. We explore a performance portable approachfor fusion plasma turbulence simulation code employing kineticmodel, namely the GYSELA code. For this purpose, we extract thekey features of GYSELA such as the high dimensionality and the semi-Lagrangian scheme, and encapsulate them into a mini-application whichsolves the similar but a simplied Vlasov-Poisson system. We implementthe mini-app with a mixed OpenACC/OpenMP and Kokkos implementation,where we suppress unnecessary duplications of code lines. Fora reference case with the problem size of 1284, the Skylake (Kokkos),Nvidia Tesla P100 (OpenACC), and P100 (Kokkos) versions achieve anacceleration of 1.45, 12...
International audienceGyrokinetic simulations lead to huge computational needs. Up to now, the semi-...
Multiscale simulation involving slow transport and fast turbulent timescales is one amongstthree key...
Recent high performance computing architectures come with more and more cores on a greater number of...
Performance portability is considered to be an inevitable requirementin the exascale era. We explore...
WACCPD 2019: International Workshop on Accelerator Programming Using Directivesisbn 978-3-030-49943-...
International audienceThis paper presents the performance portable implementation of a kinetic plasm...
With the appearance of the heterogeneous platform OpenPower,many-core accelerator devices have been ...
International audienceThe current generation of the Xeon Phi Knights Landing (KNL) processor provide...
International audienceModeling turbulent transport is a major goal in order to predict confinement p...
International audienceModeling turbulent transport is a major goal in order to predict confinement p...
Modeling turbulent transport is a major goal in order to predict confinement performance in a tokama...
Modeling turbulent transport is a major goal in order to predict confinement performance in a tokama...
This paper reports on an in-depth evaluation of the performance portability frameworks Kokkos and RA...
プラズマ乱流の運動論モデルは、4次元以上の高次元性と各次元の低い解像度に特徴付けられる。これらのコードを、性能可搬性を維持しつつ、GPUによって加速するため、OpenACCとKokkosなどのフレーム...
The goal of the extreme scale plasma turbulence studies described in this paper is to expedite the d...
International audienceGyrokinetic simulations lead to huge computational needs. Up to now, the semi-...
Multiscale simulation involving slow transport and fast turbulent timescales is one amongstthree key...
Recent high performance computing architectures come with more and more cores on a greater number of...
Performance portability is considered to be an inevitable requirementin the exascale era. We explore...
WACCPD 2019: International Workshop on Accelerator Programming Using Directivesisbn 978-3-030-49943-...
International audienceThis paper presents the performance portable implementation of a kinetic plasm...
With the appearance of the heterogeneous platform OpenPower,many-core accelerator devices have been ...
International audienceThe current generation of the Xeon Phi Knights Landing (KNL) processor provide...
International audienceModeling turbulent transport is a major goal in order to predict confinement p...
International audienceModeling turbulent transport is a major goal in order to predict confinement p...
Modeling turbulent transport is a major goal in order to predict confinement performance in a tokama...
Modeling turbulent transport is a major goal in order to predict confinement performance in a tokama...
This paper reports on an in-depth evaluation of the performance portability frameworks Kokkos and RA...
プラズマ乱流の運動論モデルは、4次元以上の高次元性と各次元の低い解像度に特徴付けられる。これらのコードを、性能可搬性を維持しつつ、GPUによって加速するため、OpenACCとKokkosなどのフレーム...
The goal of the extreme scale plasma turbulence studies described in this paper is to expedite the d...
International audienceGyrokinetic simulations lead to huge computational needs. Up to now, the semi-...
Multiscale simulation involving slow transport and fast turbulent timescales is one amongstthree key...
Recent high performance computing architectures come with more and more cores on a greater number of...