The University of Otago is a Sun Microsystems OpenSPARC Centre of Excellence. As the first Centre selected outside of the United States this is a significant achievement for the university.
The relationship between Sun Microsystems and the University recognises the University's expertise in computer architecture, networking and parallel computing. A cross-departmental team at Otago is working on a number of projects of interest to Sun Microsystems, including the Virtual Aggregated Processor (VAP). VAP is based on building virtualisation software designed to fully utilise the power of the new generation multi-core computer chips. The development of tools within VAP is intended to substantially reduce the burden on programmers by using parallel programming techniques and has huge commercial potential.
The formalisation of the relationship ensures ongoing interaction throughout a global network of researchers, programmers, engineers, technologists, and industry representatives within both Sun Microsystems and the other universities aligned with Sun. These universities are currently: the University of California, Santa Cruz; University of Texas, Austin; University of Michigan, Ann Arbor; University of Illinois, Urbana-Champaign; Carnegie Mellon University; Stanford University; Peking University; and the Rv-Vlsi Design Center.
For more information: http://www.opensparc.net/edu/centers-of-excellence.html
A View from Sweden on Multicore
I took advantage of being in Europe for the ICT 2008 Conference to visit counterparts at Ericsson and ABB in Sweden. I met them earlier in the year at the Multicore Expo in Santa Clara and both are trying to understand the implication of the multicore revolution for their business.
Coming from a university it was very interesting to get their commercial view. Both companies are taking a wait and see attitude, particularly in regard to parallel programming models that will deal with their legacy sequential code. Pretty much they are looking for industry standards to emerge before they make key decisions; they certainly see a fragmented market and worry that the companies are too small to guarantee support in their product life cycles.
However, their ability to wait for a clear solution to emerge is quite different. Martin and his colleagues at Ericsson Baseband research are dealing with technologies, such as the coming 4G LTE mobile standard, that require a growth in processing power greater than Moore's Law can provide. They have to move quickly. On the other hand, Tommy's group at ABB Corporate Research typically deals with a 15 to 20-year product lifecycle. They have more time to wait, but will only make one choice for each solution in that time.
Users, vendors and researchers are all a bit up in the air.
Further information about Multicore computing at Otago
Systems Research Group
Download an explanation of the Virtual Aggregated Processor for multicore computing (584k in PDF format)