Jeff Bier is “The Man” when it comes to DSP. His company, BDTi, bills itself as “The most trusted source of analysis, advice, and engineering for embedded processing technology and applications” and that’s not brag, just fact. The company cut its teeth on DSP-related stuff but has branched out into more diverse embedded topics. The company puts out an excellent newsletter called “Inside DSP” and this week Bier published an opinion piece on licensable cores. (See “Jeff Bier’s Impulse Response—The Rise of Licensable Cores”.) In this article, Bier discusses the industry’s increasing reliance on licensable processor cores. He writes:
“It’s one thing for a company making application-specific SoCs—which often incorporate numerous processor cores—to choose a licensable core rather than an in-house design… It’s quite another thing for a company making processor chips to adopt a licensable core rather than an in-house design; processor companies have traditionally differentiated their products in large part based on the superiority of their processor instruction set architectures and microarchitectures.”
In this paragraph, Bier is noting the increasing use of processor IP cores not just in foundry-centric SoC design but also for chips designed in house by and for sale by mainstream semiconductor vendors—specifically says Bier: “Microchip, Texas Instruments, Freescale, NXP, and STMicro.” For decades, microprocessor and microcontroller manufacturers stood squarely on the superiority of their instruction sets and processor architectures. Somehow, that approach to marketing no longer works very effectively.
Why is this happening? What’s causing the shift? I think Bier hits the nail squarely on the head when he writes:
“…I believe that most chip companies are thinking about the bigger picture when they make the decision to adopt a shared, licensable core rather than an in-house design. As processors are becoming more powerful—which of course is happening with $1 microcontrollers as well as $500 multi-core DSPs—the applications that run on them are becoming larger and more complex… to field these larger, more complex applications, system designers must work at a higher level of abstraction.”
It’s the applications. They are far more complex and far more common than ever. Products like Apple’s iPhone, iPod, and iPad have really set the bar but you can also see the flowering of apps-driven design in a wide range of consumer products such as flat-panel HDTVs, set-top boxes, and Blu-ray players.
Of course, that higher level of abstraction is a central tenet of EDA360. Design teams and project schedules do not scale with SoC complexity. If they did, the economics of SoC development would get all out of whack and we’d see the number of SoC design starts falling.
Oh, wait…that’s exactly what we are seeing.
One of the key factors driving the decline in SoC design starts is the immense and rising project costs for SoCs and the bedrock beneath these rising costs is skyrocketing complexity. We in the electronics industry have faced this problem again and again. Eventually, system-level designs become so complex that the engineering art must radically change to accommodate the complexity.
We’ve seen this happen in the transition from discrete transistors to the earliest medium-scale ICs starting in the mid 1960s. (Some of us still remember when Texas Instruments’ 7400-series TTL chips took over the world and every design engineer kept the burnt-orange TI TTL catalog close at hand.) We saw it again with the adoption of microprocessors and microcontrollers during the latter half of the 1970s, which became ubiquitous in the 1980s. We saw it with the adoption of gate arrays and then ASICs and SoCs during the 1990s as custom silicon became the only way to economically implement desired functions. At each step, engineering teams adopted radical new technologies to build ever-more-complex systems.
At its core, EDA360 recognizes this shift—the one happening right now. EDA360’s tenets highlight the pressure building as system-level complexity outstrips existing design strategies and forces us to reconsider end-to-end system design in the face of an apps-driven world.
Bier says it well when he writes “Shifting engineering investments away from reinventing-the-wheel tasks like developing a competitive-but-unexceptional processor core and basic software development infrastructure, chip vendors can invest in more innovative—and arguably more valuable—areas.”
Very good advice for any design team.