ANTAREX: Monitoring, Compilation and Autotuning Approach for Energy-Efficient HPC Systems

Europe/Prague
VŠB - Technical University of Ostrava, IT4Innovations building

VŠB - Technical University of Ostrava, IT4Innovations building

IT4Innovations National Supercomputing Center, VŠB-Technical University of Ostrava, Studentská 6231/1B, 708 33 Ostrava–Poruba, Czech Republic
Cristina Silvano (Polimi), Jan Martinovič (IT4Innovations), João MP Cardoso (Univ. of Porto), Kateřina Slaninová (IT4Innovations)
Description

Energy-efficient heterogeneous supercomputing architectures need to be coupled with a radically new software stack capable of exploiting the benefits offered by the heterogeneity at all the different levels (supercomputer, job, node) to meet the scalability and energy efficiency required by Exascale supercomputers. In this school we will start by the background and introduction to relevant topics and then will focus on the ANTAREX approach to solve some of these challenging problems by proposing a disruptive holistic approach spanning all the decision layers composing the supercomputer software stack and exploiting effectively the full system capabilities (including heterogeneity and energy management).

School participants will understand the main challenges and how the ANTAREX project provides a breakthrough approach to express application self-adaptivity at design-time and to runtime manage and autotune applications for green and heterogenous High Performance Computing (HPC) systems.

The School is specially focused on hands-on-approaches based on representative benchmarks and on the tools provided by the ANTAREX project.

Furthermore, and also very relevant, the topics addressed are specially suited for the next generation of software engineers, software developers and performance engineers (for both HPC and Cloud Computing, but also for high-performance embedded computing). The optimization of applications for heterogeneous distributed computing environments is a relevant topic in general, and we believe that organising the event in cooperation with the ANTAREX project will be a beneficial addition to the PRACE training programme.

Organizing: IT4Innovations national supercomputing center and ANTAREX project, with support of  PRACE (Partnership for Advanced Computing in Europe)

 

IT4InnovationsAntarex Project

PRACE

Slides
Travel and Accommodation Information
Participants
  • Daniel Bonczek
  • Davide Gadioli
  • Emanuele Vitali
  • František Kolovský
  • Gianluca Palermo
  • Jakub Beránek
  • Jan Chalabala
  • Jan Martinovic
  • Jan Zapletal
  • Joana Bastos
  • João Cardoso
  • Katerina Slaninova
  • Leonardo Arcari
  • Lubomir Riha
  • Martin Golasowski
  • Martin Satora
  • Martin Šurkovský
  • Nina Špániková
  • Ondřej Jakl
Support contact
  • Monday, 22 October
    • 12:00 12:30
      Registration
    • 12:30 14:00
      Lunch 1h 30m
    • 14:00 14:30
      Opening session and the role of PRACE
    • 14:30 15:30
      Introduction to Supercomputing Centers and HPC Infrastructures, Visit to IT4Innovations Supercomputing Center
    • 15:30 16:00
      Coffee Break 30m
    • 16:00 17:00
      Introduction to Parallel and Distributed Computing at the IT4I Supercomputing Center
    • 17:00 18:00
      Introduction to Monitoring, Runtime Autotuning, DSL + Source to Source Compilation
    • 19:00 21:00
      Dinner at Stračena Pub (walking distance from the training venue) 2h
  • Tuesday, 23 October
    • 09:15 09:30
      Opening session
    • 09:30 10:00
      Overview of the ANTAREX Approach
    • 10:00 11:00
      DSL and Source to Source Compilation: hands-on-approach and the Clava+LARA approach
    • 11:00 11:30
      Coffee Break 30m
    • 11:30 12:30
      Runtime Autotuning: hands-on-approach and the mArgot approach
    • 12:30 14:00
      Lunch 1h 30m
    • 14:00 15:00
      Energy-Efficiency Run-time: hands-on-approach and the COUNTDOWN approach
    • 15:00 16:15
      Use of the ANTAREX toolflow and experiments with representative benchmarks and ANTAREX Use Cases
    • 16:15 16:45
      Coffee Break 30m
    • 16:45 17:45
      Panel about Open Challenges, Exascale opportunities and challenges, Heterogeneity and the use of Hardware Accelerators, Machine Learning, etc.
    • 17:45 18:00
      Closing session
    • 18:00 20:00
      Dinner 2h