The Evil Doctor

Building an SoC is looking more like a summer blockbuster movie with a real-world villain.

popularity

I’ve always been a fan of superhero movies. I would say the Terminator series is the last time I really liked Arnold Schwarzenegger. I bet I’m not alone in that opinion. I think it’s terrific when downtrodden bands of X-Men use their strange powers to defeat evil.

The summer blockbuster season is in full swing with movies like Green Lantern and Captain America. A good time will be had by all. At the heart of these movies, there is a basic premise of good triumphing over evil. But what sets the true blockbuster apart from the also-rans (besides great special effects) is a story that goes beyond the traditional formula—something that digs deeper and takes a fresh approach.

So I’ve got superheroes on the brain today. But the superhero topic I’m thinking about goes beyond the traditional approaches. I’ve got my own “superhero” screenplay for the EDA industry. As I see it, there is a villain right under our nose that needs defeating. It’s one that has the power to destroy life as we know it. That villain is the resistance to change and complacency. Continuing with our superhero thread, let’s call this foe Dr. Sameway.

Dr. Sameway would have you believe that you can build your next SoC the same way you did it before. Just use the same point tools with the same scripts. Maybe add another server or two, a few more engineers perhaps, and you’re good to go. Many of us see the folly in this. Not too long ago, Atrenta conducted a survey regarding the challenges of timing closure. That survey found that a majority of respondents reported they had trouble closing timing on their last design because of faulty timing constraints. What would they do differently next time? The majority answer was to try harder with the same tools. Dr. Sameway won that round.

The battle with Dr. Sameway rages a lot these days around IP re-use: 80% or more of your chip consists of third-party or legacy IP. Original circuit design is becoming a boutique art form. How is all this IP verified as complete, correct and compliant with the intended application? In most cases, the user tries it and sees what happens during synthesis, place and route. They “hope” for the best. Our sales VP always says that “hope is not a good strategy.” He’s right.

So how do we defeat the insidious evil of complacency? How do we finally neutralize Dr. Sameway? I firmly believe a new and fresh approach to design methodology is needed. An approach that is more evolutionary than revolutionary, but one that clearly puts more focus on the process of design much earlier than before. The concept of focusing more effort on “getting it right” earlier in the design process is not new. It’s been around for a while, but the complexity of current SoCs has now made it mandatory. There are many names for the idea. I like SoC Realization. To stay with our theme, let’s call this superhero The Realizer. Can The Realizer defeat Dr. Sameway? I think he must, or all of us are doomed to a life of missed deadlines, over budget projects and poor growth.

Figure1

The concepts of SoC Realization have been well publicized. Google the term and have at it. Some companies have embraced the term (like us). Some are in denial. Others fear The Realizer because they believe he was invented by an evil empire. At the end of the day, it’s all hype. The realities of the marketplace will decide whether The Realizer will defeat Dr. Sameway. I’m not sure you will ever see this movie at the box office. If I had the budget, I would film it. But maybe our marketing dollars can be spent on other activities that will defeat Dr. Sameway. Stay tuned, we’ll see. And enjoy all those summer blockbuster movies while you’re at it.

 


Tags: