WE DECIDED TO LISTEN to the AMD Gang of Seven (above) to see if we could separate out the blizzard of reviews of the 2700+ and 2800+ -- future products from the chip firm ? from what actually was going to happen.
Thankfully, the usual suspects in the shape of Mike Kanellos, Mark Hachman, Nathan Brookwood, Jack Robertson and the others were there to keep the Gang of Seven honest, with some good questions about the Future Processors (FP).
Pat Moorhead, who kindly lent us a power cable at this year's CeBIT conference in Hangover, Germany, and is the VP of corporate marketing, claimed during one part of the presentation that Intel had admitted that pure clock speed was not good enough for future generations of processor.
This, no doubt, related to some information that we teased out of Intel at its recent developer forum in San Jose, and also relates to the Banias notebook chip, which launches next year at 1.30GHz before it climbs to 2GHz by the end of the year.
Moorhead told the assembly: "End users care about how their software works with their hardware. Right now we have a cryptic outdated measurement of performance.
"The buzzwords whether they're FSBs, frequencies or this or that, the general mainstream of consumers are still very confused by that".
And then this: "For close on two years... we've been on a mission to expose the Megahertz Myth.... Megahertz alone as a performance measure... is a misleading indicator of computing performance.
"Our competition has recently admitted the limitation of using megahertz alone and said it has to find an alternative way of measuring performance."
We've got an odd idea that perhaps it was our good selves at the Other Plaice that pointed out several years ago that there was a Megahertz Myth. AMD, at that time, was happy to join in the race to 1GHz, as we recall.
Anyway, we fired off an e-missive to Intel to find out what its position on the MM was.
Here's what Intel told us this fine London morning: "At IDF, we talked about Hyperthreading in the context that "Hyper-Threading Technology changes the landscape of processor design by going beyond GHz to improve processor performance" and hence system performance - along with Intel lead initiatives like USB2, DVI, S-ATA, AGP.
"Alternatively in the mobile world, beyond performance (which is still very important), mobility, form factor and connectivity are key elements in assessing the relevance of a specific platform to an individuals needs."
One of the Gang of Seven, Mark Rein of Epic Games, who apparently like the chap from Via is a logo, actually blamed the press for perpetuating the MM, pouncing on a question from dear old Jack Robertson of EB News on what the actual clock speeds of the FutureChips are (they're 2.167GHz and 2.25GHz, by the way).
Rein said: "Why do you ask that question?". He then went on to say: "The TPI, as AMD calls it is a pretty accurate measure in terms of competitive performance with Intel processors. Megahertz, gigahertz is just a meaningless statistic. The more you propagate this in the press the more you build monuments to all of this old stuff which is isn't relevant any more".
We notice that Via's Timothy Chen, who is also apparently a logo, didn't immediately jump in and say: "Oh, we at Via are going to introduce TPI for our Cyrix processors."
Nathan Brookwood asked about the new auditors for AMD's TPI initiative. As Kanellos revealed on news.com a few days ago, the auditors work for the accounting arm of PWC, and not for the division, Monday, that's going to Big Blue.
And he winkled out this highly entertaining tidbit from Ed Ellett ? who as you can see from the pic above ? is not a logo.
"PWC employees stand behind people in the benchmark labs taking digital photos of benchmarks that are on the screen."
Intel must be a bit concerned though. Ellett, although he wouldn't talk about Dell specifically, said that every significant PC manufacturer was talking to AMD about its Hammer family, and talking meaningfully.
Hybrid Mode
