Chinese high-performance computing (HPC) efforts used to be elusive and somewhat suspect to outsiders, except for the recent rise of Chinese supercomputers to the head of the TOP500 list of the most powerful systems (http://www.top500.org/). This first survey is a definitive guide to the subject.
The text contains dense statistical information, but the message is clear: government policymakers confuse HPC with mass production systems, and consequently push for supercomputer performance way beyond the requirements of the internal market.
The survey shows how the Chinese have dealt with indigenous computer design problems. It covers 60 years of computing in China, discussing national grants and policies, associations, research topics, and domestic vendors.
According to the data, the primary scientific fields that have benefited from HPC in China are numerical weather prediction and seismology. Typical industry users rarely require more than 64 central processing units (CPUs), partly because of software licensing limitations and scalability problems. The authors correctly identify that HPC software is a weak link in China, and they acknowledge the urgent need to establish a national software suite to have in mind when designing supercomputers.
Besides the bold observations on the state of affairs in Chinese HPC, this survey is most likely going to become a classic reference when discussing the subject of HPC and its applications in China.