The ANTAREX Approach to Autotuning and Adaptivity for Energy Efficient HPC Systems

Cristina Silvano (1) , Giovanni Agosta (1) , Stefano Cherubin (1) , Davide Gadioli (1) , Gianluca Palermo (1) , Andrea Bartolini (2) , Luca Benini (2) , Jan Martinovič (3) , Martin Palkovič (3) , Kateřina Slaninová (3) , João Bispo (4) , João M. P. Cardoso (4) , Rui Abreu (4) , Pedro Pinto (4) , Carlo Cavazzoni (5) , Nico Sanna (5) , Andrea R. Beccari (6) , Radim Cmar (7) , Erven Rohou (8)

1 DEIB - Dipartimento di Elettronica, Informazione e Bioingegneria
2 D-INFK - Department of Computer Science [ETH Zürich]
3 IT4Innovations - National Supercomputing Center [Ostrava]
4 FEUP - Faculdade de Engenharia da Universidade do Porto
5 CINECA [Bologna]
6 Dompé S.p.A. [Milano]
7 Sygic companies [El Segundo]
8 PACAP - Pushing Architecture and Compilation for Application Performance

The ANTAREX project aims at expressing the application self-adaptivity through a Domain Specific Language (DSL) and to run-time manage and autotune applications for green and heterogeneous High Performance Computing (HPC) systems up to Exas-ale. The DSL approach allows the definition of energy-efficiency, performance, and adaptivity strategies as well as their enforcement at runtime through application autotuning and resource and power management. We show through a mini-app extracted from one of the project application use cases some initial exploration of application precision tuning by means enabled by the DSL.

Mots clés: High Performance Computing; Autotuning; Adaptivity; DSL; Compilers; Energy Efficiency

Domaines: Autre [cs.OH]

Cristina Silvano, Giovanni Agosta, Stefano Cherubin, Davide Gadioli, Gianluca Palermo, et al.. The ANTAREX Approach to Autotuning and Adaptivity for Energy Efficient HPC Systems. ACM International Conference on Computing Frontiers 2016, May 2016, Como, Italy. ⟨10.1145/2903150.2903470⟩. ⟨hal-01341826⟩