Toward Future Computer Entertainment Systems

Abstract

There are two elements in real-timeliness that a human being can intuitively sense. One is the continuity of motion that a human being can cognitively feel to be natural, and the other is response time between action and reaction. Correspondingly, of the applications that have advanced the concept of real-time computing is computer entertainment systems which originally started as computer games. Real-timeliness of computer entertainment systems must be quick enough to match the speed of the response time of the player. Lack of both processing power and data-transfer rate in achieving this level of real-timeliness using general-purpose microprocessors have motivated the development of a new breed of more-powerful processors built on a new architecture. In computer entertainment systems, since the hardware is normally fixed for several years, once its specifications are determined, there is an inclination to seek the most-advanced technology within the future roadmap, typically that of three years ahead, and one generation ahead in the semiconductor-fabrication process. As a consequence, a system must start off by utilizing large-size chipsets at launch, but, during its life cycle, it goes through two generations of semiconductor fabrication processes for downsizing, as well as progressive integration of chips to reduce manufacturing costs, and to enable mass production. Today, more than 40 million computer entertainment systems are shipped in a year, and are becoming a strong leading power in spearheading advancement in semiconductor technology and in creating demand. In the future of real-time computing, massive assembly of "Parallel computing over the network" to execute vast amounts of computation, and "Vision System" that recognize the real world, in real-time, from a vast number of sensors over the network, will lead the next era in real-time computing