Adnan Hamid is CEO of Breker Verification Systems and the inventor of its core technology. He has more than 20 years of experience in functional verification automation. Prior to founding Breker in 2003, he managed AMD’s System Logic Division and led its verification team. Hamid worked at Cadence Design Systems, serving a system-level verification expert. He holds nine patents in test case generation and synthesis. He received Bachelor of Science degrees in Electrical Engineering and Computer Science from Princeton University, and an MBA from the University of Texas at Austin.
As you look ahead to the next few years, which embedded technologies and applications present the most interesting opportunities?
HAMID: We focus on embedded systems within System-on-Chip (SoC) devices, and the industry is hitting a wall in its ability to assemble and verify such systems. Companies are discovering that it is no longer possible just to plug together well-verified IP blocks and produce a working SoC. Complete data flows and realistic user scenarios must be verified at the full-chip level. SoC verification products such as our TrekSoC product, for example, automatically generate self-verifying C test cases to run in the embedded processors before taping out the chip. The resultant test cases rigorously stress a design, more than traditional simulation test benches or handwritten embedded diagnostics. It is critical to get this verification done because the cost of finding and fixing a bug increases tenfold for each stage later in the development process. Worst case, the process of refabricating an SoC can run into the millions, and much more if lost market opportunity is taken into account. The ability to get an embedded SoC right the first time is both a big challenge for the industry and an opportunity for Breker.
What are the largest obstacles to innovation in the embedded realm, and how should those challenges be solved?
HAMID: The verification challenge is an obstacle to producing the size and complexity of chips demanded by the most innovative projects. But the sheer complexity of SoC designs is daunting as well. You can’t innovate what you can’t imagine. Some of these chips are so huge that it’s hard for the development teams to understand all the different modes of operation, the interactions, side effects, and unintended behaviors. One way to manage embedded SoC complexity is to represent the design in the form of a graph-based scenario model. This looks like the sort of dataflow diagram that engineers draw all the time, but formalized enough so that TrekSoC can use them to generate the self-verifying test cases for verification. In this way, the developers can see behavior they didn’t expect, and even discover parallelism they didn’t know they had. Of course, this approach finds bugs, but it can also offer deep insight into designs so complex that it’s hard for any one person to understand it all.
In which market segment and geographic area do you foresee the fastest growth for embedded products?
HAMID: Consumer products of all kinds are the biggest drivers of the electronics industry as a whole, and embedded products in particular. Smartphones and tablets contain incredibly complex SoC devices unimaginable just a few years ago. The electronic content of cars continues to grow dramatically. The “Internet of Things” will connect many more kinds of devices, and not necessarily with simple chips. Multiple heterogeneous processors are the norm in almost anything connected to the cellular network or wireless Ethernet. What’s interesting is that this is happening everywhere, not just in places with traditionally better computer/Internet access. Cellular technology has become the choice for developing countries because it’s cheaper to install towers than to wire every home and business. Our consumers’ consumers are everywhere, but increasingly our consumers – the SoC developers – are everywhere too. Despite the challenges, the demands of the market and availability of engineering talent have made this a much smaller world than it used to be.
How does a company stay on the cusp of innovation, rather than just following the embedded crowd?
HAMID: Three things are essential for innovation. The first, and most obvious, is to hire the best people and set high expectations for them. Embedded developers today have to be really smart, moving seamlessly between hardware and software and from IP block to full SoC. Keeping them challenged and motivated comes naturally when pushing the boundaries of technology and products. The second is to keep up on academic and industrial research, leveraging it whenever possible. Today’s exotic proposal may well be tomorrow’s mainstream solution. Finally, find the best, smartest development teams and work closely with them. We work side by side with people developing the biggest, baddest chips in the world. That’s a lot of fun, but it means that we are constantly being pushed and challenged to evolve our products at a faster pace. We have no choice: We must innovate or die.