Intel power tech catches Transmeta. In, er, 6 months – IBM
ThinkPad exec lays it on, over and in the vicinity of the line
Somewhere within the Las Vegas Convention Centre there is a large, well-appointed room staffed by more IBM marketing people than you'd previously thought could possibly exist, clad in identical IBM shirts of a quality far higher than they've ever given you. Unlike the operatives of various other companies you might care to name, their role is not to spin, but to schedule - because Big Blue's execs are in town, dispensing wisdom in 15 minute bites to the handily cab-ranked press.
And so it was this year, we deduce,* with IBM Mobile Systems marketing director Leo Suarez being one of the prime candidates pitching yesterday. In at least two directions.
Suarez was bound to get asked about the Transmeta ThinkPad cancellation, so he was bound to have his story straight when he talked to ZDNet and CNet. Or was he? According to ZD, Suarez lashed out at Intel for focussing almost exclusively on the GHz wars while forgetting about low power CPUs for notebooks. But when Transmeta arrived on the scene, Intel "woke up." So Transmeta's power management was good, right? But according to CNet, IBM dumped Transmeta because its power management wasn't good enough. Right...
IBM demoed a Transmeta-based ThinkPad 240 at PC Expo in June, with a view to shipping it by the end of the year. Instead, whilst presumably still being undyingly grateful for Transmeta's wake-up call, it canned the project in favour of an Intel-based equivalent a few weeks back.
The Intel-based equivalent will use the forthcoming 500MHz mobile Pentium III, and you can see how brilliantly Intel's low power development and IBM's ThinkPad roadmappers are on track by the fact that the not-the-Transmeta ThinkPad 240 will be out in the middle of next year, six months later than IBM had intended for the yes-the-Transmeta variant.
But Leo clearly gives good, original back-to-back interview, and the message he conveyed to CNet was slightly different. Here, he stresses that it was the battery life problem that led to the Transmeta cancellation. Apparently IBM was aiming for seven hours battery life, and would have gone ahead if it had achieved eight (no, we don't understand that either). But the prototypes actually ran for less than six hours on a single charge. Apparently this just wasn't enough to make it worthwhile going ahead with the project. For perspective, IBM's current 240 machines have a claimed battery life of 4.2 hours, while the battery life of The Register's two year old ThinkPad 600s is now down to approx 20 minutes.
Leo argues that as Intel products now give comparable battery life to Transmeta ones, IBM didn't need to go to the expense of breaking in a new brand. Which is where the songsheets from the two interviews start to converge - "switching off Intel," he tells ZD, results in "a lot of consternation" among "large enterprises."
So here's the bottom line. IBM cancelled a project that would have allowed it to ship machines with a minimum of 50 per cent improved battery life (according to its own figures) over its current range of 240s, and switched over to another project which it says will give it similar endurance, but in another six months. Meanwhile Transmeta has promised to ship a morphing upgrade that can be flashed onto existing Crusoes next year. This will allow either performance or efficiency improvements, so presumably Suarez's equivalence with a built in six month timewarp won't be equivalent by mid-2001 after all.
But although IBM says battery life was the issue, it also points to the expense and inconvenience of not using Intel chips. Now, do we see a deep structure meaning here? ®
* 30 minutes to find the room, 20 minutes noting the shirts have eaten all the food, while the previous interview overruns, and 15 minutes of you and a Scandinavian journalist you never met before ("you don't clash so we didn't think you'd mind") being told nothing much interesting. So we don't do this stuff anymore, ourselves.