Sunday, April 21, 2024

IBM: Making Moves Beyond Silicon

Datamation content and product recommendations are editorially independent. We may make money when you click on links to our partners. Learn More.

IBM and its joint semiconductor development partners are set to begin producing semiconductors using the same breakthrough technologies that made Intel’s Penryn line of chips a success.

The process is called high-k/metal gate manufacturing and it replaces the silicon dioxide in chips that gave the Silicon Valley its nickname. By using the element hafnium instead of silicon, Intel was able to shrink its processors to 45 nanometers (nm) while at the same time improving performance and reducing heat.

Intel (NASDAQ: INTC) said it began work on this new process back in 1995 because it saw the limits of silicon-based processors coming. Up to now, the high-k/metal gate manufacturing process had been Intel’s competitive advantage, but this consortium spreads the wealth to a lot of semiconductor makers.

IBM (NYSE: IBM) and its partners, Chartered Semiconductor Manufacturing, Freescale, Infineon Technologies AG, Samsung Electronics Co., STMicroelectronics N.V. and Toshiba, were able to one-up Intel by going to a 32nm manufacturing process on 300 millimeter (mm) wafers in IBM’s East Fishkill, N.Y. facility.

They can begin taking orders now and expect to start delivering products in 2009, according to Mukesh Khare, senior manager for high-k/metal gate technology development at IBM Microelectronics.

Khare stressed that this is for all of IBM’s partners, including AMD (NYSE: AMD), which recently migrated to 45nm but not high-k/metal gate. “This process is very open,” he told “That means anybody can come and design processors, not just use it for fabrication, because if you have a fab-less design in 45nm and want to migrate to 32nm, the high-k process is very migratable and very compatible with conventional manufacturing.”

IBM claims the new circuits offer on average 35 percent better performance than 45nm technology circuits at the same operating voltage. It also claimed 32nm consumed between 30 to 50 per cent less power than the 45nm, again at the same voltage level.

Khare said this has application across the boards. “I think every market segment will get benefit out of it because it enables you to go to 32nm, which is a smaller process that requires less power,” he said. “It’s meant for low power, handheld and wireless apps, not necessarily high computing but applicable there as well.”

Fred Zieber of Pathfinder Research said the significance of this is the sheer number of major vendors involved, like Samsung, makes it a widespread benefit. “It’s a big step to move down to 32nm for everybody, and of course this makes it available to customers,” he told “You’ll get either a performance improvement or a power improvement, and some times a chip size improvement.”

This article was first published on

Subscribe to Data Insider

Learn the latest news and best practices about data science, big data analytics, artificial intelligence, data security, and more.

Similar articles

Get the Free Newsletter!

Subscribe to Data Insider for top news, trends & analysis

Latest Articles