Moore’s Law to Remain Undiminished – Intel Veteran

Transistor Scaling Does Not Slow Down, It Changes

by Anton Shilov
04/07/2010 | 11:09 PM

Moore’s Law was first described by Gordon Moore in 1965, forty five years ago. Since that a lot has changed in the semiconductor industry. Nowadays it is not enough to just make transistors smaller, but it is required to use new materials. Nevertheless, Mark T. Bohr, an Intel senior fellow and director of process architecture and integration, believes that Moore’s Law will remain unchanged.

 

“Moore’s Law and transistor scaling are not showing any sign of slowing down, but they are showing signs of changing, and this is very good for consumers and large data centers,” said Mr. Bohr in a short essay.

According to Mr. Bohr, the industry has long passed the time when new generations of transistor technology could be developed by simply shrinking vertical and horizontal dimensions. This “traditional” form of scaling has been replaced by an era where semiconductor makers have to continually develop new materials and new structures to provide scaling and meet expectations for density, performance and low power. The changes first occurred in 1999, when Intel had to replace aluminum interconnects with copper interconnects in its 180nm process technology and after that the company had to continually enhance its production processes with various new infusions.

Intel’s first 32nm products started shipping exactly two years after the first 45nm products, in late 2009, and the 22nm technology is on track for production readiness in late 2011, according to the head of Intel’s new process technologies team.

Even though Mr. Bohr does not outline any particular details about its Intel’s 22nm fabrication process, he reminds that the research group has a variety of novel transistor and interconnect ideas in the “pipeline”, including III-V channel materials, multi-gate transistors, 3D stacking and others.

“In the end, not all of the ideas explored in research make it to the manufacturing stage, but the good ones do,” said the director of process architecture and integration at Intel.

Nowadays it is no longer required to achieve maximum possible frequencies for microprocessors, instead, the goal has shifted to delivering higher performance – via micro-architectural improvements or core-count – combined with lower power. In fact, “power efficiency” appears to be the main scaling goal for chips used both in small handheld devices and in large data centers.

“So, the goals have shifted a bit, and the technology options are very different, but the value and excitement of driving Moore’s Law remain undiminished,” concluded Mark T. Bohr.