South African Research at the Speed of a Cheetah

The Centre for High Performance Computing delivers leading-edge information technology and expertise to the African research community — accelerated by NVIDIA InfiniBand networking.

The Centre for High Performance Computing in South Africa (CHPC) provides HPC resources and domain-specific support for large-scale science projects to public and private sector users in the South African research community. In addition, the Centre extends its resources and expertise to other African countries, including those in the Southern African Development Community and those that participate in the Square Kilometer Array project, an international effort to build the world’s largest radio telescope.

A service-oriented organization, CHPC helps research teams optimize their applications for top performance on the HPC systems they use for computational fluid dynamics, material science, chemistry, climate modeling and other data-intensive work. In these undertakings, CHPC works to build HPC expertise among a community of users that extends across the African continent.

HPC at the speed of a Cheetah

In all its work, CHPC seeks to ensure that the researchers it serves have access to the latest and greatest systems for high performance computing. That was the case in 2016 when the Centre introduced a peta-scale supercomputer with approximately 33,000 compute cores. The machine, delivered by Dell Technologies, debuted as the fastest supercomputer in Africa and the 121st fastest computer in the world, per TOP500 rankings.

Named Lengau, for the Setswana word for Cheetah, this groundbreaking system went into production with more than 1,000 Dell EMC PowerEdge servers and 5 petabytes of Dell EMC storage. And then in 2018, CHPC expanded its compute resources to include a GPU cluster consisting of 30 state-of-the-art NVIDIA GPUs.

Accelerating data with NVIDIA InfiniBand

Networking is a big piece of the HPC puzzle for CHPC, because fast application performance depends on fast I/O speeds. To meet this need, the Lengau system uses InfiniBand networking from NVIDIA, along with high-speed Dell Ethernet switches.

NVIDIA InfiniBand is built for the demands of complex workloads that require ultra-fast processing of high-resolution simulations, extreme-size datasets and highly parallelized algorithms. As these computing requirements continue to grow, NVIDIA InfiniBand — a fully off-loadable, in-network computing platform — provides the dramatic leap in performance needed to achieve unmatched performance in HPC, AI and hyperscale cloud infrastructure.

This is a critical need for CHPC. In fact, the center’s IT leaders evaluate applications based on how they perform on various network architectures. And they have found that many of those applications are sensitive to access to storage, so they always optimize the design of their cluster for the fabric they are using.

Powering advanced research

The value offered by CHPC comes to light most clearly in the diverse research projects the center supports. The center currently supports 26 universities, more than 900 research programs and 2,000-plus researchers who use the Lengau system, which runs close to 100 percent of the time.

Workloads running on the system are as diverse as the range of scientific disciplines. Researchers use Lengau for climate modeling and long-term weather forecasting, engineering and computational fluid dynamics, material sciences and energy storage systems, bioinformatics and precision medicine, and precision agriculture, among other research domains.

In addition, the Lengau supercomputer has emerged as a powerful tool for use in DNA sequencing, molecular modeling and vaccine development associated with the COVID-19 pandemic.

The great work taking place on the Lengau system is life changing. And at Dell Technologies, we’re honored to be the system provider in partnership with CHPC. And we’re proud to have CHPC in the ranks of the Dell Technologies HPC & AI Centers of Excellence. These distinguished centers, found throughout the world, provide thought leadership, test new technologies and share best practices with the community.

For a closer look at the Lengau system, see the Dell Technologies CHPC video above and/or read the case study. And to learn more about the CHPC and its research, visit the Centre for High Performance Computing.

About the Author: Janet Morss

Janet Morss previously worked at Dell Technologies, specializing in  machine learning (ML) and high performance computing (HPC) product marketing.