A year ago: StrongARM gets embedded in Intel’s labs

A lot of the StrongARM engineers may have decided they don’t want to go to work for Intel, but nevertheless Advanced Risc Machines and Intel managed to get together to assure the chip line’s future last week. Intel gets a licence to "produce, sell and enhance" StrongARM, and there’s also a cross-licensing deal between the companies. According to the release, "The agreement signals Intel’s plans to continue support for the StrongARM family of high-performance, low-power microprocessors, as well as plans for future enhancements to the product." But it doesn’t directly say what markets Intel will go for, or what kind of enhancements the company intends to make. "We believe the StrongARM processors have tremendous potential in the market, says Ron Smith, VP and general manager of Intel’s Computer Enhancement Group. "High-performance, low-power microprocessors are essential to the future of a variety of portable devices and other consumer electronics and embedded applications." This, we think, is a clue. Smith’s division at Intel handles embedded processors, and it’s heavily involved in the support circuitry for Intel designs, but although Intel has tinkered with more CPU-like embedded products, they’ve hardly been a major area of focus for the company. Note also that Smith isn’t raving on about pocket computers and smart phones, and he’s certainly not talking about Newton-class machines. Now, the odd thing here is that there has always been something of a clash of strategic visions between ARM and the StrongARM designers. StrongARM started off as a cheap, fast chip that would be good for embedded roles, but was headed off in more powerful directions – ARM customers, Psion in particular, saw it as maybe providing the horsepower for future devices that broke new ground in terms of power and functionality (or, as the techies have ruefully confessed to us, that just ran Java at a reasonable speed). But that Robin Saxby at ARM isn’t stupid. He’s always been perfectly willing to let his customers experiment with new markets, but he’s not going to believe there’s money in them until they take off. ARM’s revenue projections are largely based on far more boring embedded roles in far larger volumes. So maybe StrongARM was the loose cannon, and what Intel intends to do with StrongARM now is far more in line with ARM’s own thinking. ®