Rupert Goodwins' Diary

Friday 7/5/2004
Last year, I was talking to an Intel bod about Banias -- now Pentium M -- and the fact that it was really rather good.

"So it goes as fast as a Pentium 4?" I said.

"Yes," he said, patiently.

"And it's just the first of a family with a roadmap stretching off years into the future?" I said.

"Yes," he said, resignedly.

"And it takes much less power?" I said.

"Oh yes," he said, with a sigh. "So, why would anyone want a Pentium 4?" I said.

"Because," and here you could tell he was wistfully thinking of rolling his eyes in exasperation, "because it's a desktop chip with hyper-threading, and Banias is a mobile chip with power management. Of course."

"Of course," I said. "I see." I didn't, although the chap did go on to say that Intel liked having multiple teams working on similar products as it ensured a good flow of ideas. But then, with the pace of processor innovation slowing, why insist that the Pentium 4 gets hyper-threading and the Pentium M gets a power-managed stack? Why not combine all the goodness onto a single chip?

And now it looks as if this may be happening. With Dothan -- the new, smaller, faster Pentium-M -- reportedly due for launch next week with, our American cousins tell us, a top speed of 2GHz, the chances are good that the benchmarks will be even more embarrassing to the Pentium 4 bog-standard edition. More rumours say that Intel is jacking in its existing Xeon roadmap, reallocating the design teams and going M all the way. I phone my friendly Intel contact: "Oh, Rupert, there are lots of rumours flying around and you know we don't talk about speculation. If I were you, I'd take some of them with a pinch of salt". He neglected to say which chips deserved the salt.

By the time you read this, we may all know where to dispense the NaCl -- and we may all have several fewer code names to remember. Result!