As ChipCenter's DSP editor, how could I not attend? After all, Analog Devices coincidentally decided to hold the 2nd annual Sharc conference almost in my backyard, far away from the UK site of last year. And even though the conference was rather small by most standards -- roughly 100 attendees -- it attracted some very interesting people. I met engineers -- both current users as well as those taking a close look at the Sharc as a future architecture selection -- from as far away as India and South Africa who made special trips to attend the conference, not to speak of representatives from some top US companies.

Anyway, first some of the basics. The Sharc International DSP Conference 2000, held on Sept 11-13, was hosted by Northeastern Univ in Boston; I should also mention other cosponsors, which include the Univ of Hertfordshire, London Univ, Univ of Reading, Univ of Calgary, British Telecom, the IEEE and, of course, Analog Devices Inc.

During the three days, a number of presenters gave roughly 20 technical papers on all aspects of working with Sharcs. Attendees could also select from several workshops including several devoted to DSP mini training and optimizing Sharc performance in various applications.

I was particularly interested to hear some of the keynote addresses. The company's cofounder, Ray Stata, led off the conference on Monday morning. Rather than delve into DSP issues, which he left in the able hands of his staff, he decided to address higher-level issues the semiconductor industry faces. Specifically, he focused an economic theory based on Kondratieff Waves, which plot major economic activities in waves that rise and fall in periods of 50 or 60 years. A recent example in US economic history would be the crossover from rail traffic to air travel.

Stata says the electronics industry is enjoying a long "up" cycle thanks to the accumulative impact of technology in many fields coming together after a long gestation period. It's been 50 years since the transistor, and now the computer and communications technology is having a major impact on how we live and work. The web, for instance, is the product of thousands of individual innovations. Other factors include the remarkable decline in the unit-cost of labor. Other factors Stata believes will further fuel our road along the upward curve is that today's consumers can't seem to resist any new electronic gadgets, while world markets are growing due to the dropping of isolationist economic policies abroad and the fact that techology is more usable by most people.

He continued with a couple of questions: Are there any barriers to growth? How severe might they be? One issue that could slow our movement along the growth curve is an acute shortage of educated workers, especially in the US where the per-capita output of engineers has dropped below countries such as the UK, South Korea, Germany, Japan and Taiwan. We have achieved much of our prosperity thanks to the work of immigrant engineers, but the government limits the number we can take in. We've tried all the tricks we can domestically in terms of getting more engineers and making them more productive. It's reached the point where we must look around the world and employ talent whereever we can find it -- something that can also benefit underdeveloped countries. For example ADI and many other companies are setting up design centers around the world.

Stata also believes that we must address issues surrounding the federal R&D budget. He acknolwedges that the total number of research dollars is going up in recent years, but the bulk of that is going into the life sciences. Unfortunately, in recent years the number of R&D dollars going into electronics has dropped by more than 30%. To keep the economic boom going, we must get the government to redirect some of these funds into electronics research.

At the group luncheon we heard the next keynote speaker, Bob Conrad, VP and General Manager of ADI's DSP Division, who did focus a bit more on the DSP area. Again, though, he dealt more with the economics rather than technical aspects.

He first pointed out three major market areas for DSPs: general-purpose platforms, where you establish the foundation with parts and development tools and also get the word out. Second is the embedded DSP market, including the automotive, industrial and home appliance markets. Here's where you'll find product lines such as DSPs optimized for motion control. The third segment is the mixed-signal market with high levels of functional integration for apps such as broadband, handsets and v.90 modems. He did note, that there are areas where it makes sense not to integrate, one example being GSM chipsets whree the analog baseband and digital baseband are better off staying separate in a world of changing and evolving standards.

He next looked at ADI's DSP roadmap, which is diverging into both the high-performance and low-end directions. In the future we can look for two new instruction-set architectures: the TigerSharc static superscalar parallel architecture for high performance, and "FRIO", a code name for the joint development venture with Intel's communications division, whose chips will likely be targeted at cost-sensitive devices such as wireless handsets.

Conrad then gave some overall market numbers. The sales of all DSP devices this year should reach $5.8 billion -- and 75% of that amount is going into the communications sector. But you can see that there's still extremely large untapped markets if you make a close examination of the well-known bell shaped life-cycle curve. He breaks it down into these segments: Innovators, Early Adopters, Early Majority, Late Majority and Laggards. As you might expect from the numbers, communications is in the Early Majority stage; at the other extreme, military applications typically fall into the Laggard stage (meaning most military systems that need DSP have long had it). What's especially interesting is that consumer designs are now making the jump from Early Adopters into the Early Majority sector, while applications in instrumentation and industrial areas are yet in the Early Adopter stage. The upshot is that with respect to DSP, forecasts are for growth of 30-40% per year for the next decade and more. From Conrad's perspective, the best news is that ADI's DSP business grew by 70% last year, and this year looks even better with the firm already achieving 143% vs 51% for the rest of the market.

This conference wasn't someplace to go if you were looking for a ton of new-product introductions, but much of the discussion did center on the latest Sharc chip, the ADSP-21161, which ADI had formally announced just last week (check our our product review elsewhere on this DSP section of the ChipCenter site). In fact, another keynote speaker later in the week, DSP product line director Jerry McGuire, described that device and its place in the ADI roadmap.

Besides the pure technical sessions, the conference included a pleasant evening get-together at, appropriately enough, the New England Aquarium, an event cosponsored by ADI and Bittware, which supplies board-level products based on the Sharc architecture. We didn't see any sharks in the tanks or pools, but we did enjoy a sea-lion show (during which I heard the worst pun of the day: "If this is a sea-lion show, I wonder how many lines of C these lions have written?"). I won't identify the speaker so as to not destroy his reputation. A New England style lobster clambake topped off a pleasant evening under fair skies, with forecasts for both the weather and the markets for DSP quite encouraging.

		Click here to get your listing up.