Moore: Innovation key to keeping his Law alive
Moore's Law may continue for eight to 12 years
Follow @infoworld
"We're talking 10, plus or minus two, years for conventional scaling," Intel Chairman Emeritus Gordon Moore told attendees here at the International Solid-State Circuits Conference (ISSCC). But beyond that, technologies now under development may let developers keep up the pace.
"No exponential is forever. Your job is delaying forever,"
A key problem will be the need to narrow the minimum width of the smallest wires on a chip, which today are 90 nanometers wide in the most advanced chip manufacturing process. New approaches may be needed for widths less than about 30 nanometers, which will be reached in a few generations of about two to three years each, he said.
One promising new technology is the tri-gate transistor, in which the surface area of each transistor gate is increased to produce the equivalent of three gates for each transistor. Intel last year revealed plans to build such a transistor by the middle of the decade. As with dual-gate transistors under development at IBM, such a design should allow Intel to increase electrical current and thus the performance of chips without burning up the transistor or leaking electricity, the company said.
"Below 30 nanometers it's not clear that the conventional devices will work. Something like that thin transistor ... looks like a very realistic possibility,"
Denser processors also will require another big step in lithography, or etching of chips, which
Power consumption is another big issue as chips get denser, he added. Chips' voltage requirements are reduced with every generation, but it probably won't be possible to cut power indefinitely,
"This can't go on forever. You need at least a few hundred millivolts. I suspect something around 1 volt is going to be a limit, but I sure have been wrong on a lot of these other things that I suspected were going to be limited,"
The challenges in chip design and manufacturing are bigger than they seem and bigger than ever, agreed Nathan Brookwood, principal analyst at research company Insight 64, in Saratoga, Calif., in an interview Monday.
"Every generation requires greater investment in [research and development] and manufacturing to make it work, because the low-hanging fruit in terms of semiconductor production was harvested years ago," Brookwood said.









