Simon Davidmann: EDAC must evolve on multiple fronts
October 15th, 2015 by Peggy Aycinena
Imperas Founder & CEO Simon Davidmann has been thinking about the EDA industry for a while, and the consortium that represents it. And like a lot of observers, he thinks change is in the air. In previous blogs, I myself have predicted that EDAC will evolve to offer better representation to IP providers, but Davidmann believes changes in the consortium will be even more dramatic.
“When EDAC was started,” Davidmann said in a recent phone call, “it was about CAD tools. But design automation has evolved from schematic layout and simulation to a point where everything is focused on really big designs. Yes, IP is a fundamental part of that evolution and companies like Synopsys have made a lot of investment in IP, so EDAC has no problem including IP in its landscape.
“But real problems today and tomorrow are about dealing with large systems on chips. Something that is moving the focus in the industry to software. Chip design is no longer just about design tools and IP, it’s about systems, and the software that runs on those platforms.
“As a consortium designed to help companies in the design automation business, therefore, EDAC has to look at not just design tools and IP. It also has to look at systems and software. An emerging technology, quickly moving into the mainstream, is virtual platforms for software development. Of course, Synopsys is investing in virtual platforms – an indication of the importance of such things in the design process.
“But Synopsys also recently bought Coverity – a move that I don’t particularly understand because that’s a company doing static desktop software verification seemingly unrelated to design automation, though they do have some customers in the embedded software space.
“At Imperas, we do dynamic software verification and work solely in the embedded world. From our point of view, Imperas is far closer to offering something of importance to EDA than Coverity. However, we can take the Synopsys/Coverity move as a general sign of Synopsys’ view of the importance of software.
“I only offer these observations because, despite Synopsys’ move into static software verification, it’s dynamic software verification and the associated virtual platforms which we feel are the important next step in re-focusing EDAC on the future of design. We’ve spent a lot of time focusing on putting software into embedded products at Imperas, and I am absolutely certain this must be included in the evolution of the EDA industry and the consortium.
“EDAC must become an industry organization that looks after all of the companies within the circle of technologies involved in electronic product design. It can’t just be about unified solutions provided by the EDA companies. It has to be about the larger reality of hardware/software design today. EDAC must embrace the multiple directions that truly represent all design technology providers.”
Davidmann noted that the nature of the customers for design assistance is also evolving: “Today, companies in diverse market segments such as automotive, medical devices, and IoT are all worried about security. But I don’t believe we can rely solely on the EDA companies to understand these security concerns, and provide the tools to address them.
“A design automation industry must also now include companies who can solve security problems, again something that EDAC must consider in the evolution of its constituency.”
I asked Simon if EDAC is up to the challenge: “Does EDAC have the bandwidth to include all of the players that you suggest must be represented in the organization?”
Davdimann responded, “Large silicon IP players such as ARM and Imagination are working on solutions in these areas and they are part of EDAC, or have been part of EDAC. And all of the big three companies in EDA – Mentor, Synopsys, and Cadence – have been building internal tools for their own use that address all of these areas.
“Aart de Geus [Synopsys CEO] frequently talks about silicon place-and-route, synthesis, and verification, and then comments about all of the software running on virtual platforms. Mentor has an entire embedded software division, and obviously they’re moving in that direction. So you really have to ask yourself: What are the companies that EDAC represents?
“The answer is, those companies are made up of groups of people developing solutions for other groups of people, their customers, who are in turn trying to develop electronic products.
“It’s true, EDA companies are built on a different scale than companies in the embedded world. However, the EDA guys have channels and business relationships with people on both the consumer and embedded sides, so the EDA industry stands in the right place to build relationships on the ones they already have.
“If you license the Synopsys USB today, for instance, you also get the software drivers that come with it. Synopsys has to ensure that the software that ships with their IP is robust.
“In fact, when I was working at Virtual Chips [the company went IPO as inSilicon, and is now part of Synopsys], the reason we were acquired is because the silicon block was of no use without the software stack, and the software was way more complex that the hardware stack.
“Today, any complex SoC absolutely has to ship with the software to help you program it, evidenced by the fact that over the last 15 years there have been 20-to-30 companies that have put out chips that couldn’t be programmed, and have failed as a result.”
“EDA today,” Davidmann concluded, “is in the exact location it needs to be to make people aware of the importance and role of embedded software.
“EDAC can and should reflect that reality by aggressively expanding its membership, evolving to the point that it represents all companies involved in developing electronic products, the ones that focus on the hardware and also the ones focusing on the software component.
“The software in electronic products is becoming way more important than the hardware component. EDAC must embrace this.”
Tags: ARM, Cadence, Coverity, EDA Consortium, EDAC, Imagination Technologies, Imperas, inSilicon, Mentor Graphics, Simon Davidmann, Synopsys