Chuck, on Atiq Raza, I re-read the interview with him published in Oct98 Maximum PC. Here's a few interesting points he made:
*** on Intel - Q: Do you think Intel's monopolistic?
Of course, it is a monopoly. It owns the majority of the marketplace overwhelmingly. And it has used every capability that it has to utilize its monopolistic control over the market...
Intel doesn't reveal anything to create a level playing field. I'll even give you an example. Intel right now has a firmware hub under development and has given out the spec to the whole world under nondisclosure. A fundamental part of the PC platform standard and it's been distributed under nondisclosure! Why? This is a violation of every antitrust law! Here's another. After Intel licensed MMX to AMD, they went and told ISV's not to work with AMD because they had signed non-disclosures. We had a license, an open license, a legal license from Intel on MMX. We had the right to work with ISV's, but ISV's said: "Intel lawyers told us we can't work with anybody because we signed this nondisclosure." This is a totally monopolistic trade practice.
Q: Is that going to come up on the FTC investigation?
I hope so because I think Intel's practices are far, far more egregious than Microsoft's.
my comment - This makes Jerry Sanders sound like a Sunday School teacher.
*** on benefits of copper - Q: How much performance increase do you expect to get from copper?
One to two frequency bins. So if we're at 500 MHz, it will give us a 550 MHz and a 600 MHz.
my comment - more than enough to get from 900 MHz (already demonstrated in Al technology) to > 1 GHz. Raza also gave a vigourous defense of why and how copper interconnects improve performance and even sligtly reduce costs.
*** on the K6-3 - in general he was much less positive about it than the K7. He expected K6-3 systems to be available Q1'99.
*** on the Athlon - "My goal is to have the fastest microprocessor in the world with the K7"
*** on Athlon floating point - Q: The floating point on the K6-2 is notorious. Why haven't you been able to beat Intel's floating point performance?
"We will. The floating point in the K6-2 is not of the quality I would like it to be... With the K7 we'll make no apologies. Double precision, single precision, 3D technology exstensions as well as double-precision classical IEEE floating point. Again, the K7 will be the highest performance processor in both integer and floating point.
*** on Athlon servers - "We intended K7 to be a server product, and to take advantage of the server infrastructure provided by Alpha."
*** Athlon MHz - Q: What speeds will Athlon launch at?
Higher than 500 MHz
My comment - this promise was repeated verbatim until the August launch. Many commentators thought it meant that the fastest Athlon at launch would be 500 MHz. It actually came out at 650 MHz.
*** on Sledgehammer - Q: Is there a K8?
It's under development. There is a spec, but its too early to share it.
*** on the first GHz chip - Q: Will the K8 be the first GHz chip?
{laughs} Ahh, I don't know! Maybe we'll get to a GHz before that! my comment - yes, indeed they will!
After reading the whole thing over again, the overall impression is that the Athlon team has exceeded all expectations which existed 15 months ago, and this despite less than ideal partner support (VIA and motherboards), the Taiwan quake, Intel ramping to 600 MHz ahead of schedule and other internal problems (K6-3, K6-2 CXT core problems, etc).
I've gotta believe that if Raza left because he thought Dresden was expendable, he's changed his mind by now.
Petz |
