Watch Out For The Cliff
By Cary Chin
(to be sung to the tune of “George of the Jungle”)
Along with superstorms, a still-limping economy, and people launching missiles and mortars at each other, the headline news of the day continues to include the impending “fiscal cliff.” This is a completely artificial event that we created for ourselves more than 10 years ago, and it has been on our calendar ever since. It’s hard to even take our elected representatives seriously when they talk about the “current deadlock,” or “disingenuous proposals” that could add one more major bump in our road to economic recovery. So what have they been doing for the last 10 years? Campaigning?
If we had similarly ignored the technical issues associated with low power design a decade ago, we would currently have a strikingly similar crisis in chip design. With both static power and dynamic power spiraling out of control, OUR industry instead addressed the issues through research, creativity, cooperation, standardization. Those who didn’t see the problem or who didn’t invest in the solution simply don’t exist anymore.
In our case, to use an economic analogy, the “supply” of power simply hasn’t kept up with the “demand.” The figure below shows the recent discrepancy between battery capacity advancement and processor computing advancement. If the power efficiency of our designs had remained constant, there would be no such thing as an iPhone or a Galaxy smartphone, because we would not be able to power these devices for more than an hour on a full charge!
Instead, while battery technology inches forward at a relatively constant 8% rate, our industry’s performance record continues to reflect Moore’s Law—by taking advantage of the same shrinking geometries that threatened to stop technology advancement in its tracks due to power problems, we’ve used the leverage gained to solve the impending power dissipation problems. Our answer for moving to the 90nm and 65nm technology nodes was low-power design. At 40nm, we adopted even more complex advanced low-power design techniques, and at 28nm we’re prepared to introduce non-planar transistor technologies into the mix, which could keep things rolling forward for the next 5 to 10 years through the 10nm barrier.
Imagine if our politicians had the same ability to make decisions, solve problems, and move things forward. I recently heard one senator say, “Even if we don’t pass something in December (and we presumably fall off the fiscal cliff), we can just reduce taxes again when we return in January.” Are these guys kidding? The cost of non-performance in our industry would make the response very clear: “Don’t bother coming back in January.”
As we approach the holiday season, I hate to end this year on a negative note. Instead, I applaud all of you out there. Our industry has been working hard over the last decade to solve difficult problems rather than look for someone to blame for not having done our jobs. As a result, I’m writing this blog on an airplane on my iPhone 5 while listening to some great music, and I might just watch a movie in a bit— all with plenty of battery power. Thank you. And have you ever considered running for public office?
I hope you all had a great year, and wish you the best in 2013.
—Cary Chin is director of marketing for low-power solutions at Synopsys.