Grand Challenge Problems and HPC

11 Sep 2015 - 15:30
UCT Data Centre

How do we measure changes in sea level or assess what impact decreased carbon emissions will have on the earth? How do we determine the structure of a protein or decide which bacterial strains should make up a vaccine? In order to help find answers to these and other questions, researchers have to do hundreds of millions of computations in a short period of time. It was these so-called ‘grand challenge problems’ that drove the development of High Performance Computing (HPC) as we know it today in the 1990s. Today HPC is a must-have tool for scientific discovery.

HPC is a broad term referring to any computational activity that requires more than a single computer to execute a specific task. This includes the use of supercomputers and computer clusters. HPC offers researchers the capacity to handle and analyse enormous data sets at very high speeds.

“HPC is relevant to all fields of research,” says Michelle Kuttel, Associate Professor in the Department of Computer Science. “If you are doing research with big data sets that require complex analyses, you quickly realise your most precious resource is time.”

That the HPC facilities are located within the eResearch Centre, as a shared resource, is immensely helpful, says Kuttel“Having a dedicated IT team just makes economic sense. The researchers can focus on their core expertise and let the IT experts take care of the technology.”

The establishment of both HPC hardware and personnel services at UCT in 2010 has allowed researchers to tackle the grand challenge problems without having to sink time and financial resources into the acquisition of HPC hardware.

Climate change is undoubtedly one of the  ‘grand challenge problems’ facing the world today. The foremost technique to measure global change such as climate change or rise in sea level is through space geodesy. Prior to the arrival of Dr Ramesh Govind to UCT nearly two years ago, South Africa had no space geodesy analysis capability.

HPC and Space Geodesy at UCT

Space geodesy effectively measures global changes from space using data collected from satellites. On a daily basis, scientists from around the world collect enormous data sets and analyse those in order to measure the motion of the earth’s centre of mass, changing sea levels and changes in the earth’s gravity field, amongst other things. Govind says he will regularly compute 26 years of data at one time: a task that is impossible without HPC.

Space geodesy is not the only field in which researchers are advancing in leaps and bounds through the use of HPC. Researchers in astrophysics, health sciences and economics, to name just a few, are working to answer some of the most complex problems in their fields in part through the use of HPC.

HPC at UCT

While HPC at UCT may be largely driven by the needs of key research projects such as the space geodesy centre or the Square Kilometre Array (SKA) and astrophysics, the spin-off effects for researchers and students are significant.

Use of the HPC facilities has almost tripled in the past year as researchers and postgraduate students increasingly recognise the value of HPC to accelerate and enhance their research.

 

Story by Natalie Simon

Image by Stephen Williams

 

 

 

 

 

 

TOP