DAC 2015: Day 4

popularity

Are you ready for the self-driving car? Threats come from other cars – not necessarily hitting you but hacking you. Day four of DAC started with a keynote panel moderated by John McElroy of Blue Sky Productions and panelists included Jeff Massimilla from General Motors and Craig Smith from Theia Labs/OpenGarages.org/IATC.

“What would happen if cars started picking up viruses,” asked Anne Cirkel, DAC chair as she introduced the panel. Massimilla talked about the unique aspects of building electronics for cars and the different expectations for usability. Smith is a tinkerer who loves changing things, especially cars. “Cars are networks on wheels,” says Smith, “and this presents challenges to the automotive industry.” He says that we have to do reverse engineering in order to perform our ‘trade’. “It is ridiculous to think that you can make anything hacker proof.” A full description of the panel will appear on these virtual pages shortly.

Automotive panel

Having had no breakfast, and no pastries left in the press room by the time I arrived, I was eager to get to the Cadence lunch panel titled “Methodology and Metrics for Analog/Mixed-Signal Verification.” The subtitle provocatively asked – Madness or Marriage? Panelists included Neyaz Khan from Maxim, Andre Gunther from NXP and Paul Margozzi from Semtech and was moderated by Doctor Brian Fuller. Dr. Fuller had Steve Smith of Cadence on the couch in the hope of finding a cure to his problems. The problem was attempting to make an analog designer use digital verification techniques.

Cadence panel

Gunther commiserated with Steve and the way that digital has overtaken analog in many companies. Khan felt that the best approach may be to hide because he was one of the verification guys that Steve felt threatened by and Margozzi said that dealing with personality differences is always difficult. The discussion did contain some practical advice for Steve and the whole event was an interesting spin on the typical panel.

After several stops at booths on the floor, it was time for the panel session on verification titled “Scalable Verification: Revolution or Evolution.” The panel attempted to address the changes that are happening in both design and verification and the pressure that is being put on traditional tools and flows. At the same time, there are new, emerging efforts to design a new verification methodology that is top-down rather than bottom-up. Panelists Ali Habibi from Qualcomm, Steven Jorgensen from Hewlett-Packard, Bill Greene from ARM and Mark Glasser from nVidia all argued that the current system, while imperfect, was doing an OK job and could be extended to cover the additional needs of today. When pressed as to what one new thing would help them the most, they agreed on a specification language, something that would trigger a verification revolution.

Dinner was spent, not in the company of an EDA host, but with friends as we started to discuss what was good and not so good about DAC this year. For me, I thought there were some good changes, such as bringing a poster session onto the exhibition floor, the SKY talks in the pavilion booth and being given access to the keynote speakers after their talk. I will leave the not so good for private conversations.