The design of the “brain” of the Square Kilometer Array (SKA), the largest radio telescope in the world, has been completed

An international group of scientists led by the University of Cambridge has finished designing the “brain” of the Square Kilometer Array (SKA), the largest radio telescope in the world. Once completed, the SKA will allow astronomers to monitor the sky in unprecedented detail and probe the entire sky much faster than any other system currently in existence.

SKA’s Science Data Processor (SDP) consortium has completed its engineering design work, marking the end of five years of work to design one of two supercomputers that will process the huge amounts of data produced by SKA’s telescopes.

The SDP consortium, led by the University of Cambridge, has designed the elements that together will form the SKA “brain.” SDP is the second processing stage for the masses of digitized astronomical signals collected by the telescope receivers. In total, nearly 40 institutions in 11 countries have taken part.

The UK Government, through the Science and Technology Facilities Council (STFC), has committed £100 million to the construction of the SKA and the SKA headquarters as a major member of the project. The global headquarters of the SKA Organization is located in the UK at Jodrell Bank, home of the iconic Lovell Telescope.

“It has been a real pleasure to work with an international team of experts from the radio astronomy and high performance computing industry,” said Maurizio Miccolis, SDP Project Manager for the SKA Organization. “We have worked with almost all SKA countries to make this happen, which shows how difficult it is to do what we are trying to do.”

The consortium’s role was to design computer hardware platforms, software and algorithms needed to process scientific data from the CSP (Central Signal Processor) into scientific data products.

“The SDP is the place where data becomes information,” reports Rosie Bolton, Data Center Scientist for the SKA organization. “This is where we begin to make sense of the data and produce detailed astronomical images of the sky.”

To do this, SDP will have to assimilate the data and move it through the data reduction pipelines at staggering speeds, and then form data packets that will be copied and distributed to a global network of regional centers to which scientists around the world will be accessible. The SDP itself will consist of two supercomputers, one in Cape Town, South Africa, and one in Perth, Australia.

“We estimate the total processing power of SDP to be about 250 PFlop, 25 percent faster than at the IBM Summit, the fastest supercomputer in the world,” says Miccolis. “In total, up to 600 petabytes of data will be distributed worldwide each year by SDP – enough to fill over a million average laptops.”

In addition, due to the enormous amount of data flowing through SDP (about 5 Tb/s, or 100,000 times faster than the global average broadband speed expected in 2022), you will need to make your own near-real-time decisions about what is noise and what is worth preserving.

The team has also designed SDP so that it can detect and remove man-made radio frequency interference (RFI), such as from satellites and other sources, from data. “By pushing what is technologically feasible and developing new software and architecture for our HPC needs, we also create opportunities to develop applications in other fields,” says Miccolis.

High-performance computing plays an increasingly important role in promoting research in fields such as weather forecasting, climate research, drug development and many others, where state-of-the-art modeling and simulations are essential.

Professor Paul Alexander, leader of the Consortium headed by the Cavendish Laboratory in Cambridge, says the following: “I would like to thank everyone involved in the consortium for their hard work over the years. Designing this supercomputer would not have been possible without such international collaboration.”

Kathy Bennet

Leave a Reply

Your email address will not be published. Required fields are marked *