Intel's public airing yesterday of its next-generation 'Ivy Bridge' technology marks a double first for the chip giant: the production-ready debut of its 22-nanometre manufacturing process, and its 3D 'Tri-Gate' 3D transistors - a combination that promisies gains in both power and efficiency. But why both - and why now?
"This structure will provide an increase in our ability to continue scaling Moore's law," Intel Senior Vice President Bill Holt pronounced in a live broadcast from the chip giant's home in Palo Alto, referring to the dictum by Intel founder Gordon Moore that processing power will roughly double every 18 months.
Holt claimed the technology would see Intel adhering to Moore's performance curve - itself as much a way of managing the public's expectations as it is a predictor of technological development - by "at least another two years".
The benefits of the new Ivy Bridge technology are twofold: by shifting to the 22nm process, Intel can pack almost twice the number transistors into the same area, enabling a claimed performance hike of up to 37 per cent compared to the existing 32nm processor technology - or, crucially, to create smaller chips with the same processing power.
That performance boost comes coupled with the greater efficiency of the 3D design - which, Intel claims, allows a reduction in active power consumption of more than 50 per cent. Lower power consumption means longer battery life. In short, Intel has launched the next salvo in its battle with ARM for the hearts and minds of mobile device manufacturers.
Demonstrating the new technology in London to thinq_ Intel showed off what it claimed were the first two Ivy Bridge-based machines in Europe: a dual-core server based on an existing Sandy Bridge reference motherboard, and a 2GHz dual-core laptop. But in Palo Alto, Executive Vice President and General Manager of Intel's Architecture Group, Dadi Perlmutter, was quick to confirm that the technology would be applied across the company's product range:
"You can use Ivy Bridge to create everything from exo-flops supercomputers to low-voltage, low-power devices like smartphones and tablets."
Perlmutter confirmed that the new technology would be going into full production by the end of the year, with products appearing in the first quarter of 2012 - so we can reckon on seeing a rash of Ivy Bridge products at CES in January next year.
Journalists were also told to expect the technology to migrate swiftly to Intel's Atom mobile chips, with the company being able to 'shrink' the technology more quickly than it had in previous generations of chip technology.
Asked when the public could expect to see smartphones using Intel's new 3D processor technology, Perlmutter said that the company had a date - but wasn't prepared to share it.
That was pretty much indicative of the dance that Intel played throughout its announcement, playing the launch as a historic turning point in the evolution of processor technology - but the reality of the announcement was a commercial one. And the elephant in the room was ARM, the British-based company whose processor designs currently dominate the mobile device market.
Asked how the new Ivy Bridge chips' performance compared to that of ARM-based chips, Perlmutter was similarly reticent - but said tests had proved "extremely competitive". The same, he claimed, would be true of the price - with manufacturing costs on the new 22nm chips being only two or three per cent greater than previous generations.
So-called '3D' processors are hardly new - Intel previewed a tentative Tri-Gate design as early as 2002, and concedes that competitors are also working on similar technology, but Perlmutter claimed the company "expects a three-year lead" from Ivy Bridge, with rivals only planning to implement fin-based transistors - known as FinFET - in the 14nm manufacturing process, which is still some time off.
It's a head start that Intel sorely needs. While the company is moving its previously mobile-only, low-power Atom processor into both servers and desktop PCs, the chip has been synonymous with the development of the netbook - and recent figures indicate that market is on the slide, replaced by tablet devices.
And there's the rub: while Intel stalwart Dell has employed Atom processors in designs such as the hybrid Inspiron Duo, and a few Atoms have cropped up in mobile Windows-based tablets from Lenovo, ViewSonic and a welter of no-name manufacturers, the chip is very much an also-ran in a tablet market currently dominated by Apple's ARM-based iPad series.
Add to that Microsoft's announcement in January that its forthcoming Windows operating system - the mainstay of Intel's x86 processor architecture - would support ARM processors, and it was clear that Intel would have to pull something out of the bag. And with Ivy Bridge, it undoubtedly has.
Question is, is it enough?