Yes, this is about AMD, again, but I didn’t come up with the title, Digitimes did
Digitimes isn’t normally known for such curt words, so what brought that on?
Well, the author had just finished a two-part interview with an AMD exec. That could drive anyone mad.
More to the point, what set the other off was AMD talking about next year’s product rather than this year’s answer to Santa Rosa.
So far, the RS690 motherboards are beginnning to show up in OEM boxes, but the 65nm chips that are supposed to show up with them haven’t.
They’re supposed to show up in a month or so, but let’s face it, AMD hasn’t been meeting too many deadlines lately.
And that’s the problem. You just can’t talk about tommorow, you have to walk today, too. If all you do is talk about tomorrow, but you neither talk nor walk today, or what you said yesterday doesn’t apply today, pretty soon, nobody believes the talk, and you’re worse off than if you hadn’t said anything to begin with.
It’s becoming obvious that even among the enthusiast press, the propaganda is starting to become counterproductive, this outburst being just the latest example of this.
You can’t just tell people what you want to tell them, you also have to tell them what they want to know.
Let’s give Barcelona a break today, and talk about something else that will likely be important 12-18 months from now.
In the interview with the AMD exec linked above, there was this exchange:
Q: What about the future? When will you be moving to 45nm and will you be introducing new process technologies such as high-k?
A: We’re planning our 45nm transition in 2008 and you’ll see 45nm versions of Barcelona then. I don’t know what we’re releasing in terms of next generation process technology at this time, sorry.
They don’t know? This is a pretty big change, you know, not like deciding where you’re going to eat lunch.
Now unless you think the director of the server/workstation division of AMD is saying, “I’m about as much in the loop as the executive washroom attendants,” what he’s saying (and other AMD comments have indicated the same) is that AMD doesn’t know/hasn’t decided/hasn’t been told by IBM how it’s going to make processors that are supposed to debut just a little more than a year from now.
This is especially disturbing since IBM and AMD made some announcements about this about six months ago.
Those who have followed AMD a long time may be getting a sickening feeling of deja vu, and might predict that due to cost pressure, AMD might gamble that they can get away with not using extraadvanced/expensive fab tech the next go-round, and if that gamble doesn’t work, can you say, “Delay?”
Just like they gambled with 80nm R600 GPUs, and lost. Just like they gambled on single-channel memory with the initial Hammers, and lost. Just like they gambled with one less metal layer on the initial Thoroughbred, and lost, and many others over the years.
AMD does a lot of do-overs, and maybe that’s the reason for all the secrecy, to cut down the awareness of the do-overs.
Yes, Intel’s had a few situations like that, too over the years, Prescott being the most recent.
But Intel’s never been in the financial situation AMD is currently in, and they really can’t afford lots of do-overs and not delivering.
It’s odd, but the real problem with AMD is FUD, and the big AMD FUD generator isn’t Intel, but . . . AMD.
Talk about shooting oneself in the foot.
Tags: Systems & Components