Building and Auto-Tuning Computing Kernels: Experimenting with Boast and Starpu in the Gysela Code
ESAIM. Proceedings, Tome 63 (2018), pp. 152-178
Cet article a éte moissonné depuis la source EDP Sciences
Modeling turbulent transport is a major goal in order to predict confinement performance in a tokamak plasma. The gyrokinetic framework considers a computational domain in five dimensions to look at kinetic issues in a plasma; this leads to huge computational needs. Therefore, optimization of the code is an especially important aspect, especially since coprocessors and complex manycore architectures are foreseen as building blocks for Exascale systems. This project aims to evaluate the applicability of two auto-tuning approaches with the BOAST and StarPU tools on the GYSELA code in order to circumvent performance portability issues. A specific computation intensive kernel is considered in order to evaluate the benefit of these methods. StarPU enables to match the performance and even sometimes outperform the hand-optimized version of the code while leaving scheduling choices to an automated process. BOAST on the other hand reveals to be well suited to get a gain in terms of execution time on four architectures. Speedups in-between 1.9 and 5.7 are obtained on a cornerstone computation intensive kernel.
Affiliations des auteurs :
Julien Bigot 1 ; Virginie Grandgirard 2 ; Guillaume Latu 2 ; Jean-Francois Mehaut 3 ; Luís Felipe Millani 3 ; Chantal Passeron 2 ; Steven Quinito Masnada 3 ; Jérôme Richard 4 ; Brice Videau 5
@article{EP_2018_63_a7,
author = {Julien Bigot and Virginie Grandgirard and Guillaume Latu and Jean-Francois Mehaut and Lu{\'\i}s Felipe Millani and Chantal Passeron and Steven Quinito Masnada and J\'er\^ome Richard and Brice Videau},
title = {Building and {Auto-Tuning} {Computing} {Kernels:} {Experimenting} with {Boast} and {Starpu} in the {Gysela} {Code}},
journal = {ESAIM. Proceedings},
pages = {152--178},
year = {2018},
volume = {63},
doi = {10.1051/proc/201863152},
language = {en},
url = {http://geodesic.mathdoc.fr/articles/10.1051/proc/201863152/}
}
TY - JOUR AU - Julien Bigot AU - Virginie Grandgirard AU - Guillaume Latu AU - Jean-Francois Mehaut AU - Luís Felipe Millani AU - Chantal Passeron AU - Steven Quinito Masnada AU - Jérôme Richard AU - Brice Videau TI - Building and Auto-Tuning Computing Kernels: Experimenting with Boast and Starpu in the Gysela Code JO - ESAIM. Proceedings PY - 2018 SP - 152 EP - 178 VL - 63 UR - http://geodesic.mathdoc.fr/articles/10.1051/proc/201863152/ DO - 10.1051/proc/201863152 LA - en ID - EP_2018_63_a7 ER -
%0 Journal Article %A Julien Bigot %A Virginie Grandgirard %A Guillaume Latu %A Jean-Francois Mehaut %A Luís Felipe Millani %A Chantal Passeron %A Steven Quinito Masnada %A Jérôme Richard %A Brice Videau %T Building and Auto-Tuning Computing Kernels: Experimenting with Boast and Starpu in the Gysela Code %J ESAIM. Proceedings %D 2018 %P 152-178 %V 63 %U http://geodesic.mathdoc.fr/articles/10.1051/proc/201863152/ %R 10.1051/proc/201863152 %G en %F EP_2018_63_a7
Julien Bigot; Virginie Grandgirard; Guillaume Latu; Jean-Francois Mehaut; Luís Felipe Millani; Chantal Passeron; Steven Quinito Masnada; Jérôme Richard; Brice Videau. Building and Auto-Tuning Computing Kernels: Experimenting with Boast and Starpu in the Gysela Code. ESAIM. Proceedings, Tome 63 (2018), pp. 152-178. doi: 10.1051/proc/201863152
Cité par Sources :