DAC Day One

Reporter’s Notebook: Observations and insights from the DAC floor, conference rooms, and behind closed doors.

popularity

The Design Automation Confeence got off to a roaring start today and the Synopsys breakfast and keynote were standing room only.

The Synopsys breakfast brought together foundry (Samsung), IP developer (Arm) and tool provider (Synopsys) to talk about the growing requirements of ecosystems and partnership in order to make new processes available for production usage. Perhaps the most surprising aspect of the talk was that while there was a big focus on 14nm and finFETs, there was also the message that older noes are alive and well — and that will be the case for the considerable future. New 28nm processes are being created and there are many cases where 90nm and even 180nm still make sense. It is not good enough to just develop tools that are targeted at the latest nodes. They have to be able to effectively map cutting edge designs into all sizes, processes and technologies.

The DAC organizers talked about the new way in which they are creating tracks which has started to put more of a focus on applications rather than just the EDA technologies which have been the traditional focus of DAC. IP, automotive, embedded software, security and design tracks complement the tools and EDA track, with all but design having their own keynotes and a day in which they are highlighted. Of course, DAC does not last long enough for each track to have its own day, and today sees IP and management tracks sharing the stage.

The keynote was given by the CEO of imagination, whose talk contained a few key points. Hossein Yassaie talked about the lead time for when they need to start developing a new piece of IP and its eventual inclusion in end-user products, which can be six to seven years. With that kind of lead time you have to do a lot of work to understand the market, project where the technology is going, and the needs that it will have. Then you can start to design for those goals. The other point he made, which seconded a message from Gary Smith’s Sunday night talk is that you must have deep domain knowledge if you hope to be successful. Gary seemed to imply that without it you would be doomed to be a developer of IP that would be designed out as quickly as possible. Later in the day semiconductor engineering sat down with Peter McGuiness, director of marketing for imagination, and the subject explore in more detail. Expect to see that interview published in the next few days.