I came to DAC 2009 looking for the industry trends in ESL, because as we all know by now, ESL can mean many things to many people. At today's lunch panel titled "Are SystemC and TLM-Driven Design Ready to Replace RTL?" moderator Mark Johnstone put up a picture of an elephant and listed about 10 things ESL might be, depending on the angle from which you look at the elephant. I don't remember them all, but they were all quite familiar.
Two pieces caught my eye before DAC that I was looking to confirm or deny. One was titled "What should EDA do next?" (a catchy title for those of us wishing to remain employed). In the article, Paul McLellan talks about the importance of the embedded software and the Virtual Platform. The second was a fragment from "I Sense A Tremor In The Force" by John Cooley showing some evidence of actual usage of High Level Synthesis. The phrase that caught my eye was "not that la-la fruity virtual prototyping crap". John is clearly focused on hardware implementation.
My sense from DAC so far is that High Level Synthesis and Virtual Platform are the top two opportunities in ESL. Sure, some people will be always be interested in the other 8+ things ESL might be good for, but these two seem to be rising to the top as the most pressing issues with the biggest demand from users.
For High Level Synthesis, users are taking a hard look at the technology including not only quality of results, but also overall productivity. The good news is that Cadence is in position to provide both High Level Synthesis and Virtual Platform solutions. When users start to see the synergy between a common source that can be used for both purposes and get concrete information on a complete flow on how to do both at the same time, the benefits look much more compelling than looking at either High Level Synthesis or Virtual Platform alone. There are still some kinks to be worked out, and only time will tell, but things are trending positive. If the vision of a common starting point for both High Level Synthesis and Virtual Platform cannot be worked out, then users are stuck with the same issues they have now, a separate flow for hardware implementation and abstract simulation.
Not only is Cadence is a good position to address both opportunities, but both are actually managed by the same person, Mac, and all of the people are in the same organization, almost as if by design.