IBM GATE-CRASHES chip world, boldly exclaims: 'We've cracked the 7nm barrier'

When can you buy new wonder processors? Don't ask

IBM has reached a major milestone in its computer chip R&D efforts – by shrinking transistor gates down to 7nm, far below today's proportions.

On Thursday, Big Blue announced that it has successfully produced test chips with functional transistors using a 7nm (nanometer) process technology, and claim an industry first.

Generally speaking, the smaller you can shrink the transistor gates on a chip, the more efficient the chip will be.

Various physical effects kick in as you shrink the gate size: for one thing, the dynamic power consumption of the transistors drops, but the static consumption increases as they leak more current. With the right techniques to mitigate this leakage – such as sticking gate fins on the transistors – as you shrink the process size, you can run the chips at higher clock speeds with lower power consumption, which in turn leads to cost savings, generally speaking.

Today's best chips are fabbed using process technologies at scales of 22nm or 14nm. Moore's law suggests the number of transistors on dense integrated circuits tends to double every couple of years, so 10nm and 7nm processes are expected to be the next natural rungs in the ladder.

But while the industry's current chippery techniques have got us to 14nm, going down to 7nm is another matter. At that scale, it gets extremely difficult to control the electrons flowing through the transistors. And the next step – getting beyond 7nm – is going to be even more challenging.

As IBM explained in an email to The Register: "Pursuing such small dimensions through conventional processes has degraded chip performance and negated the expected benefits of scaling – higher performance, less cost and lower power requirements."

That's why, 12 months ago, IBM announced that it would invest $3bn over five years to develop new chipmaking technologies that have a shot of breaking the 7nm barrier.

Working with partners GlobalFoundries, Samsung, and State University of New York's SUNY Polytechnic Institute Colleges of Nanoscale Science and Engineering, IBM has now said it's getting closer to achieving such a breakthrough.

Specifically, what Big Blue and its partners have managed to do is use silicon-germanium (SiGe) channel material, rather than plain silicon, to produce features at the 7nm scale.

Sponsored: Minds Mastering Machines - Call for papers now open

Next page: Broadwell? Pfft

Biting the hand that feeds IT © 1998–2018