High performance computing

The CHPC is one of three primary pillars of the national cyber infrastructure intervention supported by the Department of Science and Technology (DST) and managed by the Meraka Institute of the CSIR.

The South African National Research Network (SANReN) complements the CHPC through the provision of high-speed, high-bandwidth connectivity, and the very large databases (VLDB) provide the effective curation of a variety of extraordinarily large databases.

The CHPC recently awarded the University of the Witwatersrand just over R2-million to be used over the next three years for its innovative cyber technology project.

The team, led by Prof. Steve Damelin of the Wits School of Computational and Applied Mathematics and Prof. Micheal Sears of the Wits School of Computer Science, will investigate hyperspectral imaging which refers to the collection and processing of information from the ultraviolet to thermal region of electromagnetic spectrum.

One of the projects will be to apply hyperspectral imaging using satellite technology to create patterns of data that will enable geologists to study rock materials and minerals that could assist in prospecting for valuable minerals and other resources. Prof. Damelin says this is where mathematics comes in. “The team will look at producing algorithms to compress and represent high dimensional spaces on two dimensional computer screens and to produce data that is easy to access and understand. This is an example of a ‘cyber loop initiative’ which brings real life problems to the table to be interrogated and solved mathematically by students and researchers in applied mathematics.”

The fastest supercomputer in Africa

Housed at the CSIR in Cape Town, the CHPC’s SUN Constellation system is the fastest supercomputer in Africa and amongst the fastest in the world as rated by the top500.org list of top 500 super computers. The machine is made up of the latest Intel Nehalem 8-core processing units with a state-of-the-art on-rack water cooling system.

Using its over 2000 processors and 400 TB of storage, along with its smaller siblings (the 256 processor 2 TB shared memory SUN M9000 and the 384 core SUN Harpertown cluster) the CHPC aims to enhance significant research, address grand challenges, and grow computational research into a viable mode alongside experiment and theory across all academic disciplines.

The CHPC supports several high level research projects including computational space physics, energy storage and the Meerkat and SKA projects

Space physics

The computational space physics project constructs, links and expands numerical models to simulate the transport and acceleration of charged particles (cosmic rays) from their creation in the galaxy to their arrival on Earth. Calculations are required for the acceleration of these particles at astrophysical shocks in supernova remnants, their propagation in the galaxy, and the transport in our local turbulent astrosphere (the heliosphere).

The emphasis is on particle propagation, the study of the formation, flow profiles, magnetic field and geometry of super nova remnants and the heliosphere itself. These simulations will help to test different theories, and to explain recent measurements from various spacecraft and large telescopes such as the H.E.S.S. gamma-ray telescope in Namibia.

The end results will be applicable to studies of the influence of cosmic rays and space climate on the climatic environment of Earth, as well as long duration missions to Mars, and to working environments on the moon and Mars.

Energy storage

High energy-density solid-state lithium-ion batteries are increasingly used in many applications, ranging from low-power mobile telephones, electronic notebooks, through storage of solar and renewable energy, to high-power vehicle traction. Insecurity in the supply of fossil fuels and the necessity for substantially minimising environmental pollution have recently accelerated usage of such batteries in electric vehicles (EV) and hybrid electric vehicles.

The main purpose of the project is to use computational modelling methods in the enhancement of cost-effectiveness, energy and power density, charge/discharge time, and number of charge/discharge cycles of high power rechargeable batteries. The best approach to any battery design and optimisation programme, which is lacking in most instances, is to align it directly to applications it is intended for. In the current study, battery performance evaluation results derived from the EV and renewable energy (Eskom) development projects, will provide valuable insights on changes that are needed in battery chemistry and configuration in order to provide targeted outputs.

MeerKat and SKA

This project addresses electromagnetic computer simulation of key elements of the proposed Square Kilometre Array (SKA).

The world’s radio astronomy community is working together on the concept ofthe SKA – the largest and most sensitive radio telescope ever. It is likely to consist of a myriad dishes, each 10 – 15 m in diameter. Special antenna tiles in the core of the array will form a “radio fish eye lens” for all sky monitoring at low frequencies. This will allow many independent simultaneous observations. The joint receiving area of all these dishes and panels will add up to approximately 1-million m2. The SKA will require super-fast data transport networks to support its need for increasingly powerful computing. South Africa and Australia are the only countries on the shortlist to sitet this mega-telescope.

The MeerKAT, another very important South African project with a budgeted cost of R800-million, is considered to be the largest government-funded science project since the early 1990s. The primary elements in both the SKA and KAT are the receiving antennas.

HPC in business

There are already a number of organisations in the business world that deploy high performance computing (HPC) solutions to deal with big data complexities and the challenges of today’s environments.

While for most companies conventional computer power is adequate, there are instances when access to HPC capabilities can increase efficiency, improve response times, enhance decision-making and help manage the increasing complexity of data.

As technology develops and costs decrease, HPC is increasingly likely to extend from the research environment into also solving business problems. It is certainly an interesting development to watch in the future.

High performance computing in SA << Comments and views

Source: EngineerIT

Latest news

Partner Content

Show comments


Share this article
High performance computing