March 07, 2011
By the end of the year, Chinese supercomputers will stop using imported processors and will instead rely on homegrown chips, according to an article in People's Daily Online. Billed as the world's fastest supercomputer, China's Tianhe-1A system, which gets its CPUs and GPUs from US vendors Intel and NVIDIA respectively, could be the country's last to use foreign-built components. Going forward, supercomputers will employ China's domestically-developed Loongson series of microchips, also known as the Godson processor.
On Saturday, National People's Congress Deputy Hu Weiwu, lead developer of the Loongson series of microchips at the Chinese Academy of Sciences (CAS), revealed that the institute's upcoming "Dawning 6000" supercomputer will use 10,000 of the Loongson series chips to achieve a petaflop of processing power.
Implementing these chips into supercomputers is a national priority for China, the article reports. The Institute of Computing Technology of CAS, the Jiangnan Institute of Computing Technology and the National University of Defense Technology (NUDT) are all on schedule to start development of Godson processor-based systems this year. In addition to defense workloads, the machines will be used for science applications such as geology, meteorology and medicine.
Despite the significance of these developments, the Chinese HPC industry is still figuring out how to create applications that can fully leverage such massive manycore systems. The article also cites a lack of software engineers with supercomputing expertise. Hu explains:
"We have enough supercomputers in China but still can't fully utilise them. There are lots of scientific questions waiting for answers from supercomputer calculation. But we still need good algorithm and good data collection to make it work."
Because of the software problem, when the Dawning 6000 goes online, researchers will only be able to extract about 10 percent of its potential. Even with the limited usage, the cost to power the system will not be cheap. Hu anticipates yearly energy costs of 10 million yuan, or $1.5 million.
While concerns that these native chips will cut into the US microprocessor market are understandable, there may be time to mount a technological counterstrike. According to Hu, the ability to mass market the chips is still a ways away, but the intent to do so is clear:
"It still needs another decade before China-made chips meet the needs of the domestic market. Hopefully after two decades, we will be able to sell our China-made CPUs to the US just like we are selling clothes and shoes."
Full story at People's Daily Online
10/30/2013 | Cray, DDN, Mellanox, NetApp, ScaleMP, Supermicro, Xyratex | Creating data is easy… the challenge is getting it to the right place to make use of it. This paper discusses fresh solutions that can directly increase I/O efficiency, and the applications of these solutions to current, and new technology infrastructures.
10/01/2013 | IBM | A new trend is developing in the HPC space that is also affecting enterprise computing productivity with the arrival of “ultra-dense” hyper-scale servers.
Ken Claffey, SVP and General Manager at Xyratex, presents ClusterStor at the Vendor Showdown at ISC13 in Leipzig, Germany.
Join HPCwire Editor Nicole Hemsoth and Dr. David Bader from Georgia Tech as they take center stage on opening night at Atlanta's first Big Data Kick Off Week, filmed in front of a live audience. Nicole and David look at the evolution of HPC, today's big data challenges, discuss real world solutions, and reveal their predictions. Exactly what does the future holds for HPC?