​Square Kilometre Array software prototype runs on Chinese supercomputer

The International Centre for Radio Astronomy Research (ICRAR), a joint venture between Curtin University and The University of Western Australia, has announced that software prototyped for the Square Kilometre Array (SKA) has been tested on the Tianhe-2 supercomputer in China.

The successful deployment of the prototype science data processor execution framework was carried out on the world's second-fastest supercomputer out of the National Super Computing Center in Guangzhou, China.

The test was conducted by an international team led by professor Tao An from Shanghai Astronomical Observatory in China and professor Andreas Wicenec, head of data intensive astronomy at the ICRAR, in Western Australia, with the test forming part of the Science Data Processor work package for the SKA, led by the University of Cambridge in England.

According to the ICRAR, the execution framework provides the control and monitoring environment to execute millions of tasks, consuming and producing millions of data items on thousands of individual computers -- the scale of processing required for each SKA observation obtained within a six to 12 hour period.

The SKA has been slated to be the largest and most capable radio telescope ever constructed, with ICRAR calling it the world's largest science project, covering over 1 million square metres of collecting area.

Scheduled to begin construction in 2018, the SKA is an international project which will consist of thousands of antennas spread across the world, with central cores of operation in South Africa and Western Australia. Its central computer alone will have the processing power of about 100 million PCs.

At 50,000 times more sensitive than any other radio instrument currently in existence and powerful enough to detect very faint radio signals emitted by cosmic sources billions of light years away from Earth -- including those emitted shortly after the Big Bang, over 13 billion years ago -- the SKA is expected to help scientists answer fundamental questions about the universe and the laws of nature.

Wicenec said the complete system currently being designed will process such raw observations and convert them into a form that can be analysed by astronomers.

"It is known as the SKA Science Data Processor, or the 'brain' of the telescope", Wicenec said.

The Australian professor said the execution framework of the science data processor is "data activated", meaning individual data items are wrapped in an active piece of software that automatically triggers the applications needed to process it.

"Whenever a data item is ready, that's triggering the next task -- the task is not running idle, waiting for anything," he said.

The prototype was initially run on 500 compute nodes of the supercomputer and was then extended to 1,000 nodes. The ICRAR confirmed the next step would be to "ramp up" the number of individual items that are getting deployed and then increase the number of computer nodes to near 8,500.

The system is now running 66,000 items and the next stage will be a few million, with Wicenec hoping to push this out to between 50 million and 60 million items on 8,500 or 10,000 nodes.

As part of the federal government's AU$1.1 billion National Innovation and Science Agenda unveiled in December, AU$294 million was given to host the West Australian sector of the SKA, with the government saying the SKA will deliver significant economic, scientific, and technological benefits to the country.