Intel: “Moore’s Law Still Applies”

32-nanometer chips about to launch – first samples of 22-nanometer technology.

Archive notice: This article was originally published on September 23, 2009. Links and embedded videos are preserved as part of the historical record.

On Tuesday evening at the Intel Developer Forum (IDF), Intel not only presented the “Westmere” processor, which is manufactured with feature sizes of 32 nanometers (nm) and is already scheduled to be produced for sale from the fourth quarter of 2009. For the first time, memory test chips were also shown that had already been manufactured using 22 nm production technology. “Moore’s Law still applies at Intel,” said Intel President and CEO Paul Otellini at the IDF. From an expert perspective, of course, it was clear that Intel would achieve feature sizes of 22 nm – the more interesting question is whether the chip manufacturer can continue Moore’s Law in the longer term.

According to Moore’s Law, the number of transistors on a given area doubles every 18 to 24 months. The 22 nm technology is a step along a path whose end is foreseeable for skeptics. Market research company iSuppli, for example, said in June of this year that a point would be reached below feature sizes of 20 nanometers where manufacturing would become too expensive for mass production. Intel, by contrast, is more optimistic. “As a rule, we only talk about new production processes when we are about a year and a half away from product introduction,” Intel spokesperson Martin Strobel told news service pressetext.de. The company was confident that it could continue Moore’s Law with silicon technology for several more years. Intel’s laboratories were also researching approaches such as using other elements like hafnium or new structural forms. “Carbon nanotubes are also a major topic,” says Strobel. These are also in the research spotlight as a microchip material in academia, partly because of interesting material properties.

“Moore’s Law means that more and more functions fit on chips,” Strobel emphasizes. This can also be seen in Intel’s computing present. After the company only recently integrated some chipset tasks into the main processor in the Lynnfield chips, Westmere will for the first time integrate a graphics processor core (GPU) directly into the main processor. “This brings the flexibility of the x86 architecture and the parallelism of GPUs closer together,” says the Intel spokesperson. The chip, which is close to mass production, will be followed next year by “Sandy Bridge,” a new microarchitecture using 32 nm technology. It integrates sixth-generation GPUs into the processor and also promises accelerated floating-point calculations. At the IDF, video and 3D applications were demonstrated on an early Sandy Bridge-based system to underscore the performance of this architecture.

Topics

No Topics assigned to this article.

Article details

Published
Author
Archive
Reading time
About 3 min
Word count
435

Tags