The ANTAREX Tool Flow for Monitoring and Autotuning Energy Efficient HPC Systems

Cristina Silvano (1) , Andrea Bartolini (2) , Andrea Beccari (3) , Candida Manelfi (3) , Carlo Cavazzoni (4) , Davide Gadioli (1) , Erven Rohou (5, 6) , Gianluca Palermo (1) , Giovanni Agosta (1) , Jan Martinovič (7) , João Bispo (8) , João M. P. Cardoso (8) , Jorge Barbosa (9) , Kateřina Slaninová (7) , Luca Benini (10) , Martin Palkovič (11) , Nico Sanna (4) , Pedro Pinto (8) , Radim Cmar (12) , Ricardo Nobre (8) , Stefano Cherubin (1)

1 DEIB - Dipartimento di Elettronica, Informazione e Bioingegneria
2 D-INFK - Department of Computer Science [ETH Zürich]
3 Dompé S.p.A. [Milano]
4 CINECA [Bologna]
5 IRISA-D3 - ARCHITECTURE
6 PACAP - Pushing Architecture and Compilation for Application Performance
7 VSB - Technical University of Ostrava [Ostrava]
8 FEUP - Faculdade de Engenharia da Universidade do Porto
9 LIACC - Departamento de matematica aplicada [Porto]
10 ETH Zürich - Eidgenössische Technische Hochschule - Swiss Federal Institute of Technology [Zürich]
11 IT4Innovations - National Supercomputing Center [Ostrava]
12 Sygic companies [El Segundo]

Designing and optimizing HPC applications are difficult and complex tasks, which require mastering specialized languages and tools for performance tuning. As this is incompatible with the current trend to open HPC infrastructures to a wider range of users, the availability of more sophisticated programming languages and tools to assist and automate the design stages is crucial to provide smoothly migration paths towards novel heterogeneous HPC platforms. The ANTAREX project intends to address these issues by providing a tool flow, a DSL, and APIs to provide application's adaptivity and to runtime manage and autotune applications for heterogeneous HPC systems. Our DSL provides a separation of concerns, where analysis, runtime adaptivity, performance tuning and energy strategies are specified separately from the application functionalities with the goal to increase productivity, significantly reduce time to solution, while making possible the deployment of substantially improved implementations. This paper presents the ANTAREX tool flow and shows the impact of optimization strategies in the context of one of the ANTAREX use cases which is related to personalized drug design. We show how simple strategies, not devised by typical compilers, can substantially speedup the execution and reduce energy consumption.

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

Domaines: Autre [cs.OH]

Cristina Silvano, Andrea Bartolini, Andrea Beccari, Candida Manelfi, Carlo Cavazzoni, et al.. The ANTAREX Tool Flow for Monitoring and Autotuning Energy Efficient HPC Systems. SAMOS 2017 - International Conference on Embedded Computer Systems: Architecture, Modeling and Simulation, Jul 2017, Pythagorion, Greece. ⟨hal-01615945⟩