Concurrency has become a hot topic in the past few years because the concurrency avaible in hardware is increasing. Processor designers have reached a point where make sequential tasks take less time has become impractical or impossible -the physicl limits encountered demand enineering tradeoffs. But writting software that takes advantges of concurrency is hard, and making that software perfom as well on different CPU architectures is all but impossible. In this article, we will explore the reasons why this currently true, with specific examples, and will considerer how this evolution represents a changing paradigm that renders the traditional imperative programming model fragile and infficient.
© 2001-2024 Fundación Dialnet · Todos los derechos reservados