China has set its sights on becoming a world leader in AI, and their recent investment in building an exascale computer is a major step in that direction.
Checkout this video:
The Birth of Exascale
In 2005, Jack Dongarra, a well-known computer scientist, together with three other scientists, proposed the Exascale Computing Research Project (ECRP) to the US Department of Energy (DOE). The aim of the project was to develop a machine that could perform one exaflop: a million billion (1018) calculations per second. This is a thousand times faster than any computer that existed at the time, and would require around a thousand times more power.
The ECRP was not funded, but in 2006, the DOE launched the Frontier project, with a similar aim of developing an exascale machine by 2018. The project was led by Peter Nugent, then at Lawrence Berkeley National Laboratory (LBNL).
In 2007, China set up its own national project to develop an exascale machine, called the Chinese Exascale Computing Initiative (CECI). The CECI was initially headed by Weng Weijie from the Shanghai Supercomputer Center (SSC); however, in 2009 he was replaced by Guangwen Yang from the Chinese Academy of Sciences (CAS).
The US and China have taken different approaches to building their exascale machines. The US has focused on using graphics processing units (GPUs), while China has developed its own microprocessor architecture.