Patent Energy Shifts to Power
Battery life is critical for smartphones, tablets and Ultrabooks and as more people embrace mobile technology, power has become an innovation hotbed in the competition for vital intellectual property, according to a longtime Intel engineer who has filed 50 patents.
"The battlefront has moved," said Jim Dodd, Intel principal engineer. "It's not about performance, but performance per watt, especially when you move into phones and tablets -- low-power stuff. Every patent that saves a milliwatt is incredibly useful."
Dodd joined Intel 22 years ago and presently is part of the Visual and Parallel Computing Group based in Folsom, Calif. He filed his first patent, which improved the speed of DRAM, back in 1997.
While DRAM hasn't changed since the late-'90s, patent filing has in basic ways, according to the Intel engineer, adding that having good architecture and solid digital design are still important, "but that isn't where the patents are." Dodd cited USB 3.0 as an example of something that pushes the envelope of speed, but isn't necessarily about the digital aspects. "It's the analog, the software, the circuitry. That's where the innovation is," he said.
Like a baseball player whose debut at-bat is a walk-off home run, Dodd says of the 50 U.S. patents he's filed over two decades, that first DRAM patent was his biggest.
Issued in February 1997, Patent #5603010 improved the speed of DRAM in chipsets. "Performing Speculative System Memory Reads Prior to Decoding Device Code," as the patent is named, played to what Dodd described as "funny properties" inherent to DRAM.
"If I want to read some data out of the RAM, it depends on where that memory is stored because what you do is access a neighborhood, so to speak, of where that info is," he said. "If you want to read other data in that neighborhood it's quick. If you want to read data from a different neighborhood you're going to close that neighborhood and open another one so it's slower. The slower something is the lower the performance."
What the technology in Dodd's patent does is guess that the data is located in the same neighborhood that is already open. If wrong, a one-clock "penalty" is taken, according to the inventor, explaining that a clock is a small unit of time. "But if I'm right," Dodd said, speaking as the computer, "then I just shaved some time off of that access, and more often than not, I'm right."
The improved performance from patented technology that trims clocks off the latency gave Intel the best-performing DRAM in the industry, according to Dodd. "We had the lowest possible lead-off time for accessing memory by 30 to 40 percent over everyone else. It really was common sense what we did, really simple, and it ended up being a great feature for the 82340FX chipset," he said.
Although 16 years old, Patent #5603010 is still used today because, as Dodd said, "the same idea of doing a speculative read while you go decode the address to figure out where the data is still works."
Though Dodd hasn't received a patent since earning his milestone 50th in 2008, it's not due to a lack of innovation. "Patents are cool because it marks a milestone in what you've done, but not everything an engineer invents meets the criteria for a patent," he said. "Many of the things I've worked on are buried so deep into the chip it's not something a lawyer can say, 'Hey, you're infringing on our patent!'"
An example of such input is called a trade secret, and although they won't get your name recognized by the U.S. Patent and Trademark Office, they can give your company a competitive advantage. One trade secret of Dodd's enabled Intel to use existing testing equipment for a new 324-pin ball grid array, or BGA, it invented and licensed to the world, which prior to then used 192-pin surface-mount packaging for integrated circuits.
"I figure that this trade secret saved the company tens of millions of dollars, so its value is more than any of my patents," Dodd said.
The engineer of 47 years and 50 patents, in addition to a forgotten number of trade secrets, credits a collaborative work environment for his accomplishments. He offered some advice for any of the next generation of engineering types who thinks they can thrive in a vacuum.
"You have to be willing to exchange ideas and share problems you encounter," he said. "No one ever goes off in a corner and invents something by themselves. Even Thomas Edison didn't do that. Rarely do you ever see a patent with only one name attached to it. Creativity comes through communication -- the constant exchange of questions, which leads to ideas."