Re: Where's the THUMPERTURD Copper Wipey ?
Well Paul,
I was skeptical when AMD claimed they'd be able to ship 650MHZ parts - but they did. I was skeptical when AMD claimed they'd be able to design and ship a stable chipset - but they did. I was skeptical when AMD claimed they'd be able to ship volume at 700, then 750, then 800, and then 850 - but they did. In this morning's Chicago Tribune, just like last week, local screwdriver shops are offering 800 and 850MHZ Athlons. Best Buy has 800MHZ Athlon systems from HP on the shelf.
This is ahead of forecasts. So I have come to trust the investor relation data from AMD.
On the other hand,
Intel claimed they would be able to produce enough high speed parts to end the coppermine shortage in January - but they didn't. Then they said they'd solve the problems in February - but they didn't. Will they catch up in March? Will we have plenty of 1GHZ parts in March? Catching up doesn't mean shipping the target speeds from 6 months ago - catching up means shipping today's target speeds today. It is now March, and Dell wants $320 and an extra 30 days to substitute an 800 coppermine for a 733!
Intel is falling further and further behind.
Intel claimed they'd ship stable rambus chipsets/motherboards in September - but they didn't. Then they said they'd solved the rambus problems in December, then in January - but new problems showed up. What about the Profusion errors?
At this point, as an investor, I have a lot more confidence in "AMD PR burble" than anything from Intel.
There was a quote recently from a named Intel source giving an on the record interview about Itanium (I have no Idea where I saw it - maybe someone with a better memory than mine can help out with a link). And he made a big deal about moving the power pins off the the chips pin grid to reduce pin count. This is a chip that targets only high end systems, in very low volume (in the initial package, at least) and it sounds like one of the most focused on design goals was pin count.
I think that Intel has evolved a number metrics for internal assessment that have resulted in pernicious incentives poisoning the development process. That certain chip packaging parameters are considered more important than stability, cost, or performance seems to have been repeatedly demonstrated by the various Rambus fiascos. Now it looks like the same set of management goals is driving the Itanium/Mckinley development process.
Just my opinion, but it seems to fit the facts.
Dan |
