Following the PRACEdays14 conference in Barcelona at the end of May, the organisation has appointed a new Prace Council Chair and announced awards for papers presented by both industry and academia.
At the 13th Council Meeting of Prace – the Partnership for Advanced Computing in Europe – which was held on 3 and 4 June 2014 in Athens, Greece, Sanzio Bassini of CINECA, Italy was elected Chair, with Anwar Osseyran of SurfSARA, Netherlands elected as Vice-Chair.
Prace has also recently announced the recipients of several awards, 3 of which were presented at PRACEdays 14, and the final award, the Prace ISC Award given to the best paper at the International Supercomputing Conference (ISC) in Leipzig in June. The PRACE-ISC Award will be presented to Alexander Breuer, TU München; Alexander Heinecke, TU München; Sebastian Rettenberger; TU München; Michael Bader, TU München; Alice-Agnes Gabriel, LMU München; Christian Pelties, LMU München for their paper entitled ‘sustained petascale performance of seismic simulations with SeisSol on SuperMUC’.
The PRACEdays14 Award for Best Scientific Presentation went to Teodoro Laino of IBM Research – Zurich, Rüschlikon, Switzerland for his presentation entitled ‘shedding light on lithium/air batteries using millions of threads on the BG/Q supercomputer’. The PRACEdays14 Awards for Best Industrial Presentation was presented to Mathis Bode of RWTH Aachen University, Germany for his presentation entitled ‘high fidelity multiphase simulations studying primary breakup’.
The PRACEdays14 Award for Best Poster was given to Kannan Masilamani of University Siegen, Germany for his poster entitled ‘simulating and electrodialysis desalination process with HPC’.
Over 200 scientific and industrial participants attended the conference and its satellite events, which included an open session of the PRACE User Forum, and a Workshop on Exascale and PRACE Prototypes.
Since 2010 Prace has allocated over 8 billion core hours on 6 world-class machines with a total capacity of nearly 15 Petaflop/s, hosted in 4 EU Member States (the PRACE Hosting Members France, Germany, Italy and Spain), to 303 scientific and industrial projects from 38 countries.
The total funding of Prace amounts to €530 million over 5 years (2010 – 2015). Around 35 per cent of Prace access is given to non-hosting and non-EU countries, and awarded projects come from a diverse group of scientific and industrial domains.
The Prace Council decides on all matters of the Association. It is composed of one representative from each Member although, depending on the nature of the decisions to be taken, different voting rules apply.
Sanzio Bassini said: ‘I am looking forward to taking up this task and guiding PRACE into the next stage of its existence to let Prace 2.0 become solid and concrete. There are many challenges ahead, but I am confident that we can make PRACE grow and prosper, the foundations for Prace 2.0 having been laid by my predecessors.’ Catherine Rivière, the outgoing chair of the Prace council, said: ‘Supercomputers are part of the great human adventure and my dream for PRACE has always been to make it a sustainable and world-leading research infrastructure. Prace 2.0 will bring that dream one step closer to reality.’
The PRACE Scientific Steering Committee (SSC) selects the paper to receive the Award, which will be presented at the ISC Opening Session on Monday 23 June 2014. The winner of the PRACE ISC Award will receive sponsorship for participation in a training event, or a conference relevant to petascale computing.
On behalf of the winners, Alexander Breuer said: ‘Seismic simulations in realistic 3D Earth models require peta- or even exascale compute power to capture small-scale features of high relevance for scientific and industrial applications. In this paper, we present optimisations of SeisSol — a seismic wave propagation solver based on the Arbitrary high-order accurate DERivative (ADER) Discontinuous Galerkin method on fully adaptive, unstructured tetrahedral meshes — to run simulations under production conditions at petascale performance. Improvements cover the entire simulation chain: from an enhanced ADER time integration via highly scalable routines for mesh input up to hardware-aware optimisation of the innermost sparse-/dense-matrix kernels. Strong and weak scaling studies on the SuperMUC machine demonstrated up to 90 per cent parallel efficiency and 45 per cent floating point peak efficiency on 147k cores. For a simulation under production conditions (108 grid cells, 4.8x1010 degrees of freedom, 5 seconds simulated time), we achieved a sustained performance of 1.09 PFLOPS’.
