Clock speed is no longer the most important measure on processors prowess.
It has been supplanted by performance per watt, which addresses the greening of the chip industry. The performance bump that formerly came from cranking up clock speed is now the province of multicores. The only problem is that most software isn't good at taking advantage of multicore architectures.
To overcome that hurdle, Stanford University is partnering with Sun Microsystems, Advanced Micro Devices, Nvidia, IBM, Hewlett-Packard, and Intel to create software that will allow chips to more efficiently process many tasks at the same time, according to a report in The New York Times. The effort--dubbed the Pervasive Parallelism Lab--is expected to be announced by the group on Friday, according to the report.
The project follows similar efforts announced last month by Intel and Microsoft, which are committing a combined US$20 million to fund parallel computing research centers at the University of California at Berkeley and the University of Illinois at Urbana-Champaign.
The separate university efforts will share some approaches, but will also try different experiments, including new programming languages and hardware innovations, according to the Times.
Stanford's lab will cost US$6 million over three years and be led by electrical engineering professor Kunle Olukotun, a pioneer the idea of multicore microprocessors, according to the Times.
This article was originally a blog post on CNET News.com.
Play video
» Maximum flexibility with powerful blade technolgy
Secure the "Next-Gen SOA Infrastructure" & "Bringing SOA Value Patterns to Life" whitepapers here
There are currently no comments for this post.