Hosting Sites & Projects

Successful applicants will have the opportunity to spend two months at one of the following hosting sites, working on these projects:



CINECA  –  Consorzio Interuniversitario

Visualizing the results of the project “The way to heating the solar corona: finely resolved twisting of magnetic loops”

The PRACE Tier-0 project “The way to heating the solar corona: finely resolved twisting of magnetic loops” investigates the heating released by the magnetic reconnection of a progressively twisted Solar Coronal Loop. We propose to visualize the results obtained in this project.

Visualization of turbulent structures of the wake

The flow around a wind turbine is complex: it is possible to see the pylon and the nacelle wake, the hub and tip vortices and the wake expansion. It is important to see how the turbine operation regime affects the development of the turbulent flow structures downstreams.
The visualization of the turbulent structures of the wake are possible extracting information from the flow (iso-surfaces of pressure, vorticity, lambda2, etc) in order to see the development of the complex wake structure realizing a short clip.


EPCC  –  The University of Edinburgh
Edinburgh, United Kingdom

Dinosaur racing to demonstrate the role of HPC in simulation

What’s faster: a Tyrannosaurus, Velociraptor or Dilophosaurus? Using HPC we can find out by combining paleontology with biology to form a simulation of these ancient creatures. What’s more, this provides an engaging visual demonstration of how HPC can be applied to the sciences which is especially applicable to outreach events such as the British Science Festival. By allowing the public to design and race dinosaurs against each other, we can make a lasting impression about how simulation is the third research methodology, complementing theory and experiment.

In conjunction with a 2013 SoHPC student, we have created a visual outreach demo centred around dinosaur simulation and racing. The public can configure their own dinosaur, accurately simulate it on the previous UK national supercomputer, HECToR, and then race it against other dinosaurs to see who can create the fastest. This is an illustration of how simulation is critically important where experimentation is prohibitive or impossible and participants receive a certificate illustrating their custom creature along with its vital statistics (weight, height, speed etc.) The demo has been very successfully deployed at a number of outreach events including the British Science Festival and the EU Young Scientist awards in Prague, the same team reusing the demo at another EU PRACE outreach event in January.

Based upon this success, we would like to further develop the demo to not only keep it fresh but also enable it to run on ARCHER – the new UK supercomputing service. The research team of palaeontologists have given us a number of new, highly detailed, dinosaur models and we would both like to incorporate these into the demo and also allow for much more freedom in design dinosaurs by combining limbs of different creatures (they are currently limited to modifying the size of the limbs of an Argentinosaurus.)

As an additional project goal, we would also like to investigate using tablets to design the dinosaurs. This would allow for the public to be designing their creatures concurrently which will both improve the interactivity of the demo and increase capacity.

Visualizing Energy Usage

This project will involve extension of existing programs for visualizing energy usage in computers to consider multiple system components, systems features and allow easy interpretation of results.
There are a number of methods for gathering energy data but no standard method of visualization. Visualizing data (from log files or in near-real-time) would benefit the community and be an interesting research topic.

There are currently a few simple methods for interpreting and comparing systems but these could be expanded upon and developed to provide insight. Algorithms and metrics for performance comparison are also in their infancy and the student would be have direct input into these, with potential to have their work included in a future publication.

This work is not restricted to x86-based systems and the work will make use of multiple systems, from small embedded SoC systems to supercomputers, including different processors and accelerator cards.


JÜLICH  –  Forschungszentrum Jülich
Jülich, Germany

Simulation of physical many-body systems in MD

The simulation of physical many-body systems in MD requires the computation of the Coulomb forces and potentials. Many MD-simulation tools, like GROMACS take up substantial runtime for these computations. Modern algorithms, like the Fast Multipole Method (FMM), help to reduce the runtime without deteriorating the requested error bounds. Besides the efficiency of the algorithms the porting/optimization onto modern hardware is essential. In this project we want to extent the current FMM to work on GPUs.

Simulations of Lattice Quantum Chromodynamics

Simulations of Lattice Quantum Chromodynamics are used to study non perturbative properties of strongly interacting matter. These simulations take up a large fraction of the available supercomputing resources worldwide. The student will be involved to tune the most critical parts of a multi-grid solver recently co-developed by the proposer to the Intel XeonPhi architecture. He will test extensions of the method and thus get acquainted with High Performance Computing and Numerical Mathematics alike.


ULFME  –  University of Ljubljana, Faculty of Mechanical Engineering
Ljubljana, Slovenia

Geometrically accurate wind barrier model in numerical simulation

In previous studies, porous wind barrier in numerical simulation was modeled as a momentum sink which excludes the barrier geometrical characteristics. Exclusion of the barrier geometrical details represents a superficial level of fluid flow simulation near the barrier.

However, nowadays HPC offers massive parallelization which permits accurate porous barrier geometry model to be used in the numerical simulation. This offers an insight into the fluid flow structures through and around the barrier which depend on the barrier geometry.
Deeper understanding of the wake structures behind different barrier models and the possibilities that HPC parallelization offers will be presented in this work.

ParaView plugin for fusion data structures

3D scientific visualization in HPC environments is a topic that ranges from post-processing (on dedicated visualization clusters) to in-situ code instrumentation. Often, 3D visualization is based on multi-layered data access frameworks that need custom plugins to be developed for specific codes. SoHPC applicant will (based on LLNL VisIt plugin) work on “porting” prototype-plugin tailored for fusion data-structure model. Developed plugin will give further insight and possibility to explore deficiencies and propose architectural requirements (to Kitware) that could be of general interest of HPC and PRACE community.

The project is about programming in C++ with XSLT code generation. The aim is to extend existing program (UAL database reader plugin running under VisIt visualization) with prototype for ParaView for different meshes ranging from 1D plots (VTK/Charts) to 3D unstructured meshes.

Interfacing fusion codes is based on Consistent Physical Objects (CPOs). CPOs are standardized data structures that describe various physical aspects of fusion experiments and are designed to be suitable for use with simulation codes and experimental data. Integration with CPOs thus brings a common data model to simulations that allow to directly comparing results with experiment, use experimental data as input or mixed approaches. To facilitate change and to support different programming languages, the data structure is described by a XML schema definition (XSD) from which visualization schema in XML is generated. A basic usage of the data structure description is persistent storage in a database. Presently, plugin works, but several shortcomings were spotted within various workflows.

Applicant will work on learning of existing plugin code (2500 lines in total) and porting to ParaView including the following topics:

  1. Examples and upgrades of 3D VTK visualizations
  2. Upgrades of XSLT to visualization schema
  3. XSLT to ParaView plugin
  4. User interface for browsing databases (optional)

Target visualizations will be selected by the topic of interest and will be provided. In any case student will work closely with mentor on coding, testing and visualizations.


B  –  Technical University of Ostrava
Ostrava, Czech Republic

Car race game demonstration

The objective of this project is to develop a car race game which should demonstrate capabilities and application of HPC systems. This game will be used for popularization of HPC among young generation. In this game players will have freedom to change design of a car (size, shape) which will change their chance to win the racing game. Through this game basics of fluid dynamics (i.e. aerodynamics) and effect of drag coefficient on car performance (car speed) will be demonstrated. Students assigned to this project will collaborate to achieve this goal. One student will work on GUI and game engine. Another will develop automated CFD simulation of external aerodynamics of car for drag coefficient calculation. This student will work on whole work flow i.e. automated mesh generation and simulation using open source code OpenFOAM.

Please follow and like us:
Posted in Hosting sites & projects 2014, PRACE Summer of HPC 2014 Tagged with: , , , , , ,
Follow by Email27