Intelligent approaches to niche markets involving parallel processing in multicore chips could bring New Zealand a lucrative new industry and reputation – so optimists in a panel session concluding the Multicore World conference in Wellington recented contended.
There are “smart people” in New Zealand, who can move into an under-recognised area of innovation, acting as subcontractors to major ventures in Silicon Valley and other centres of technology expertise in the US, panel convenor Martin McKendry of conference organiser Open Parallel suggested. The US companies would supply application program interfaces and software development kits for New Zealand-based teams to work directly on specialised niches contributing to a large project.
Such a venture would be likely to operate without hope of significant return until the teams have built up experience and established a reputation, McKendry suggests; but in time it could be very lucrative.
Government’s recognition of the “high-value manufacturing” sector – a priority for the new Ministry of Business, Innovation and Employment - can only be positive for such an effort, said NZ Trade and Enterprise’s John Houlker. The proposal for an Advanced Technology Institute to be built around the Industrial Research Ltd crown research institute identified stronger connections between universities and industry. “A university’s success is currently measured by citations on [international academic] papers, not connection with industry,” Houlker said. A venture like that proposed could be a step towards turning that attitude around.
Houlker referred to the Auckland Bioengineering Institute’s Physiome Project, where mathematical models of all human organs are being created for drug testing “in silico” (by digital simulation), as evidence that New Zealand can take part in groundbreaking digital projects.
However, there was considerable scepticism about the multicore proposal, particularly from Intel’s Tim Mattson. US universities are busy building just such expertise, he told panellists and conference audience, and it is growing under the spur of intense competition between strong rivals such as the Massachusetts Institute of Technology (MIT), Harvard, Yale and Georgia Tech – rivalry which does not seem as strongly in evidence among New Zealand institutions. The Association of Computing Machinery (ACM) has a “very active programme” of reformatting the US university syllabus and this is likely to include parallel processing, either as a separate elective part of the syllabus or an addition to all aspects of computing, he said.
Gaining expertise in parallel processing “won’t be a big leap” for Kiwi students, suggested Catalyst IT director and NZRise president Don Christie, from the audience. “Innovation happens on the edge”, not in huge technology campuses, he said and New Zealand computing people are used to working on that edge.
“You don’t think [it’s a big leap],” Mattson retorted. “The leap is vast”; performance doesn’t figure highly in computer science students’ and their tutors’ thinking, he suggested.
Conference organiser Nicolás Erdödy said, speaking from the position of an entrepreneur, he is sure there are gaps in the market; identifying those and forging connections with the US companies would put the venture on a sound basis. “What I am suggesting is that we figure out some hypothesis involving an area where our work is further out than what [major US companies] are doing,” then make approaches. “I’m not begging,” he said to the US delegates, “but give us some help, please.”
Erdödy ran the common argument that such a unit could act as a core for overseas experts to gather around, given that “clever people can live where they want to,” and New Zealand is a pleasant place to live. The one drawback, he conceded, is the current inadequacy of broadband communications.