The ongoing merger of embedded and EDA
Much has been written over the last few years about how the electronic design automation (EDA) industry must reach outside of its traditional boundaries to grow. Consolidation in the semiconductor industry and the enormous costs to develop and fabricate new chips has resulted in fewer EDA customers and fewer ICs being designed. The common conclusion is that EDA will expand into the embedded systems space. There’s less speculation on this possibility in the embedded world, but it’s worth examining the arguments.
It’s natural for EDA companies to set their sights on software. They have primarily served the needs of hardware designers, a community that’s growing, but is relatively small. In contrast, there are countless millions of programmers in the world. With the introduction of “apps” for smartphones and tablets, it seemed as if everyone became a weekend programmer. This represents a potentially huge new customer base. However, the market for selling tools to this community is very small, both in terms of need and willingness to spend.
Embedded sits between EDA and the great masses of programmers. Embedded engineers often do some amount of hardware design, usually involving circuit boards and programmable devices. Embedded designs are often complex enough to require simulation, circuit analysis, and other EDA technologies. Embedded programmers work much closer to the hardware than apps developers, and often have intimate knowledge of the entire system. Serving the needs of everyone on an embedded team should be good business, especially if the Internet of Things (IoT) really takes off.
Arguably, EDA’s growth into embedded is much less than the opposite direction. No chip of any consequence is a pure hardware design anymore. Systems have always required a mix of hardware and software, of course, and microcode blurred the line decades ago. Today, the two domains are more intertwined than ever. Programmable cores of every description fill system-on-chip (SoC) designs, implementing many functions that formerly used dedicated hardware logic.
These days, embedded programmers aren’t just working on boards that contain big chips; they’re working side by side with design engineers developing those chips. Beyond writing the production software, embedded expertise is needed in simulation, emulation, FPGA prototyping, and the bring-up lab to write diagnostics for verification and validation purposes. The two industries might not have truly merged yet, but in the trenches, designers and embedded programmers are working side by side.
There are quite a few signs that EDA and embedded are moving toward being a single market. The Accellera standards organization has a working group focused on developing a “portable stimulus” standard. This will define migration from IP block to full SoC and from simulation to silicon. This goes far beyond the historical simulation testbench domain of Accellera, including generation of C test cases for SoC processors that can replace many of the diagnostics currently hand-written by embedded programmers. When available, this standard is likely to draw the hardware, software, and embedded engineers even closer together.
On the vendor side, the move into software is a key piece of the EDA360 vision that Cadence painted a few years ago. Synopsys took a huge leap into the software world, even beyond embedded, with its acquisition of Coverity last year. Mentor Graphics has been in the embedded world with its own real-time operating system (RTOS) for years. Perhaps the most obvious sign of the ongoing trend is the growing overlap in technical content between traditional EDA conferences and traditional embedded conferences.
One current example of this is the technical program for the Design and Verification (DVCon) India show in Bangalore September 10-11. Although primarily an EDA conference, there are numerous sessions that cross into embedded territory. Several talks and tutorials discuss the portable stimulus concept. Closely related to this topic are talks on graph-based verification, an abstraction that can be used to drive automatic generation of some types of production software as well as diagnostic code. Other sessions relevant to embedded engineers include pre-silicon testing of drivers, hardware-software co-verification, and IoT development challenges.
It is unclear whether the embedded and EDA worlds will completely merge, served by a single set of vendors and viewed as a single market. However, the arguments that the merger will continue to happen incrementally are strong, and supported by specific examples of changes in the industry. It will be very interesting to watch this ongoing evolution and to see how it affects the way that engineers work together to succeed on the most demanding projects.
Tom Anderson has more than 15 years of experience in EDA verification applications. Currently he is Vice President of Marketing at Breker Verification Systems. He has served as Product Management Group Director for Advanced Verification Solutions at Cadence, Technical Marketing Director in the Verification Group at Synopsys, and Vice President of Applications Engineering at 0-In Design Automation. Before moving into EDA he was Vice President of Engineering at IP pioneer Virtual Chips, following roles in ASIC design and management. Tom has presented more than 100 conference talks, published more than 200 papers and articles, and contributed to 12 books. He holds a BS in Computer Systems Engineering from the University of Massachusetts at Amherst and an MS in Electrical Engineering and Computer Science from the Massachusetts Institute of Technology (MIT).