For those who missed it, Asus announced the EEE PC at Computex. A super small 7" notebook running Linux priced at $199.
Press Release
Flash slide show
What Asus and Intel seem to be keeping mum about is what's inside. Folks are guessing it's either an A100 or A110 part. However, I'm curious. These parts are finding their way into several devices. But not one of them comes even close to this price point. Discounting the cost of not handing over a big fat cheque to Microsoft, I'm baffled how Asus managed to hit this price point unless Intel is giving Asus the parts for free.
Now I may be bullish on Intel currently, but I'm not a putz to believe Intel gives their customers free parts. So I'm wondering how difficult is it for Intel to whip out 80 wafers of Silverthorne before they actually get into mass production. I mean that's all they need considering each wafer gives them 2500 CPUs.
So if any of the tech geniuses out there know the answer, I'd be delighted to find out. Can Intel run 80 wafers? How feasible is this considering Otellini first revealed this product perhaps 3 months ago.
Subscribe to:
Post Comments (Atom)
8 comments:
Enough to kill AMD.
I thought this thing was supposed to run on Linux, not Windows.
Operating system aside, the real answer to your question can be seen by looking at battery life. Silverthorn systems are supposed to run on 1/4th the power of current systems. If this unit is built around Silverthorn, the battery life should be a give away.
"Can Intel run 80 wafers? How feasible is this considering Otellini first revealed this product perhaps 3 months ago."
80 wafers is nothing and 3 months would be doable if the wafers were hotboxed through the fab. The real problem I see with this theory is that I would think silverthorne would need much more validation after the wafers came out of a fab.
Given that it is a new product, at a minimum Intel would have to do extensive rel and burn-in testing (accelerated lifetime) Now I guess it's possible Intel has already done this, but I find it unlikely.
Another possibility is that is is run by a low voltage, low clocked Celeron that has been sitting on Intel's shelves long enough to be written off. From an an accounting perspective, it's cost is ~0~, making the chip highly profitable at any ASP.
"Another possibility is that is is run by a low voltage, low clocked Celeron that has been sitting on Intel's shelves long enough to be written off."
This seems more plausible - Intel could even have run a few lots through a fab which is primarily running chipsets (only talking 3-4 lots here).
The cost in this case would also be close to nil - as equipment is already bought and being used for other purposes (chipsets). Running 80 wafers through a fab that runs 5000 wafer starts a week would not even be noticed, and not "steal' from potential chipset capacity as it is so small qty.
Thus the only real incremental cost would be the Si wafers and packaging, both are inconsequential...
At larger quantities the capacity cost would start to matter and perhaps this is the purpose of Silverstone and the long term strategy. For short term and fairly low quantities, this would be felt as much as a fly on an elephant from a capacity perspective.
I'm not sure about Celeron. I think any Netburst part regardless of clock will not support the thermals for that form factor. And I doubt they would be writing off any Cnroe or mobile parts.
How about 5w TDP? I'm not sure what chip this is, but Banias or Dothan should be capable of it.
http://www.intel.com/pressroom/archive/releases/20040720comp.htm
Two things - package size and cost. Intel would not be writing off any mobile parts and this thing is too cheap to hit the price points of any of their existing mobile products. And none of their mobile parts except for A100/110 would really fit and those are too expensive...as far as I know.
Post a Comment