ECMWF hosted the fifteenth workshop on high performance computing in meteorology from 1 to 5 October 2012.
Description
Every second year the European Centre for Medium-Range Weather Forecasts (ECMWF) hosts a workshop on the use of high performance computing in meteorology.
The emphasis of this workshop will be on running meteorological applications at sustained teraflops performance in a production environment. Particular emphasis will be placed on the future scalability of NWP codes and the tools and development environments to facilitate this as we move towards petaflop computing.
Our aim is to provide a venue where
- Users from our Member States and around the world can report on their experience and achievements in the field of high performance computing during the last two years; plans for the future and requirements for computing power will also be presented.
- Vendors of supercomputers will have the opportunity to talk to managers and end users of meteorological computer centres about their current and future products.
- Meteorological scientists can present their achievements in the development of parallel computing techniques and algorithms, and can exchange ideas on the use of supercomputers in future research.
- Computer scientists can give an update on their efforts in providing tools which will help users to exploit the power of supercomputers in the field of meteorology.
- The challenges of creating a computer centre infrastructure for HPC can be discussed.
The workshop consisted of a limited number of presentations from invited speakers, plus a series of 20 minute contributions. The morning of the final day was reserved for an open discussion session.
Programme
Presentations
| Monday 1 October | |
|---|---|
Session 1 |
|
|
ECMWF forecasting system - research and development Erland Källén (ECMWF) |
 |
|
Performance of IFS on ECMWF's new HPCF Deborah Salmond and Peter Towers (ECMWF) |
|
Session 2 |
|
|
Data intensive supercomputing at NCAR Anke Kamrath (NCAR) |
- |
|
NCEP high performance computing acquisition and migration Allan Darling (NCEP) |
|
|
NCEP applications - HPC performance and strategies Mark Iredell (NOAA) |
|
Session 3 |
|
|
Towards zero cost IO in the Unified Model Martyn Foster (Met Office) |
|
|
HPC enabling NWP at the Canadian Meteorological Centre Bertrand Denis (Canadian Meteorological Centre) |
|
|
The CMC 15-km Operational Deterministic Global Vivian Lee (Canadian Meteorological Centre) |
|
Session 4 |
|
|
A survey of performance characteristics of NOAA's weather and climate codes across our HPC systems Craig Tierney (NOAA) |
|
|
Scalability of Elliptic solvers in numerical weather and climate-prediction Eike Mueller (University of Bath) |
|
|
Towards a scalable performance-portable software infrastructure for the Gungho Dynamical Core Rupert Ford (STFC Daresbury) |
|
| Tuesday 2 October | |
|---|---|
Session 5 |
|
|
Weather and climate simulations with BULL technology Franck Vigilant (Bull) |
|
|
A status update on HP's solutions for HPC, illustrated with examples from Research and Operational Weather Michael Lough (Hewlett-Packard Galway Ltd) |
|
|
Kepler GPU architecture and benefits to Earth System Modeling Stan Posey (NVIDIA) |
|
Session 6 |
|
|
Hybrid solutions for Meteo sites Yuichi Kojima (NEC) |
|
|
Early experiences with IBM p775 and ENDGame Paul Selwood (Met Office) |
|
|
Findings from real petascale computer systems with meteorological applications Toshiyuki Shimizu (Fujitsu) |
|
Session 7 |
|
|
Global modelling research at NOAA/ESRL Sandy MacDonald (NOAA) |
|
|
An update on fine-grain computing activities at NOAA ESRL Mark Govett (NOAA) |
|
|
Performance and scaling of the NIM Global NWP Model on GPUs Tom Henderson (NOAA) |
|
|
Experiences porting a weather forecast model to the new Many Integrated Core (MIC) architecture from Intel Jim Rosinski (NOAA) |
|
Session 8 |
|
|
Towards a scalable semi-structured data platform for COPE Anne Fouilloux (ECMWF) |
|
|
The Icosahedral Nonhydrostatic (ICON) model: Scalability on Massively Parallel Computer Architectures Florian Prill (Deutscher Wetterdienst) |
|
|
A stable and accurate Kalman filter for large scale weather data assimilation Tuomo Kauranne / Alexander Bibov (Lappeenranta University of Technology) |
|
| Wednesday 3 October | |
|---|---|
Session 9 |
|
|
RAPS introduction George Mozdzynski (ECMWF) |
|
|
The enhanced DWD-RAPS Suite - Testing computers, compilers and more? Ulrich Schaettler (Deutscher Wetterdienst) |
|
|
Parallel I/O for Scalable Earth System Modelling Luis Kornblueh (Max Planck Institute for Meteorology) |
|
|
Offloading I/O in Arome Sami Saarinen, CSC - IT Center for Science Ltd |
|
Session 10 |
|
|
Capability system with interconnect for global addressability and hardware collectives Michael Woodacre (SGI) |
|
|
Multi-scale multi-physics simulations and toward next step on the Earth Simulator Keiko Takahashi (Earth Simulator Center) |
|
|
How to overcome common performance problems in legacy climate models Jorg Behrens (DKRZ) |
|
Session 11 |
|
|
Optimisation of weather applications on power on x86 architectures Francois Thomas (IBM France) |
|
|
Climate supercomputing at NASA Goddard Space Flight Center Phillip Webster (NASA Center for Climate Simulation) |
|
|
New developments with the TotalView Debugger: Scalable messaging and support for the Intel Xeon Phi Chris Gottbrath (Rogue Wave Software Inc) |
|
|
From IFS to OOPS Yannick Tremolet / Mike Fisher (ECMWF) |
|
Session 12 |
|
|
An update on the European Network for Earth System Modeling Reinhard Budich (Max Planck Institute for Meteorology) |
|
|
A PGAS implementation by co-design of the ECMWF integrated forecasting system (IFS) George Mozdzynski (ECMWF) |
|
| Thursday 4 October | |
|---|---|
Session 13 |
|
|
Leadership Computing at the National Center for Computational Science: Transitioning to Heterogeneous Architectures Jim Hack (Oak Ridge National Laboratory) |
|
|
Adapting Numerical Weather Prediction codes to heterogeneous architectures: porting the COSMO model to GPUs Xavier Lapillonne (MeteoSwiss) |
|
Session 14 |
|
|
Management system of a vast number of operational jobs at JMA Masami Narita (Japan Meteorological Agency) |
|
|
Satisfying operational requirements for fault-tolerance of NWP ensembles George Carr (NOAA/GSD) |
|
|
Earth System Modeling at the Extreme Scale Per Nyberg (Cray Inc) |
|
Session 15 |
|
|
Progress in NWP on Intel HPC architecture at the Australian Bureau of Meteorology Robin Bowen (Australian Bureau of Meteorology) |
|
|
A multiscale non-hydrostatic atmospheric model for regional and global applications James Doyle (Naval Research Laboratory) |
|
|
The current and future US Navy Global Prediction Systems: Developments under ESPC Melinda Peng (Naval Research Laboratory) |
|
|
Meeting timeliness and throughput constraints in IASI L2 operational processing Tim Hultberg (EUMETSAT) |
|
Session 16 |
|
|
Advanced Lustre infrastructure monitoring Torben Kling-Petersen (Xyratex) |
|
|
Accelerating data management Glenn Wright (DataDirect Networks) |
|
|
High performance computing in Serbian Hydrometeorological Service Milan Dacic (Republic Hydrometeorological Service of Serbia) |
- |
| Friday 5 October | |
|---|---|
|
The global nonhydrostatic atmospheric model MPAS: Preliminary results from uniform and variable-resolution mesh tests William Skamarock (NCAR/NESL, MMM) |
|
|
Stabilizing high-resolution HARMONIE Enda O'Brien (Irish Centre for High-End Computing) |
|
|
Atmospheric soundproof model with three-directional MPI parallelization Zbigniew Piotrowski (IMGW-PIB) |
|