Interview: Intel's Don MacDonald talks Centrino

 

Gallery

Posted on 26 Sep 2003 at 16:54

MacDonald: We haven't set pricing yet, but it is not obvious that if you had a Dothan processor compared to the Banias that it would cost more.

If we want to make the transition to our 90nm process (on which the Dothan processor will be based), we have a manufacturing pressure on us to say 'make the move from 200mm wafers and .013micron to 90nm with 300mm wafers' - we have a financial incentive to go do that. The best way to do that - this is not rocket science, purely the law of supply and demand - is to say 'if you want to convert all your Banias customers to Dothan, price them at the same price'. It's a much better product, why would you buy the other one?

So that's one of the scenarios I am currently leaning towards, it doesn't have to be more expensive - 'you can get more at the same price point'.

You've already mentioned the 90nm process for Dothan, and another featured technology is its use of 'strained silicon'. Can you explain this in layman's terms?

MacDonald: I'll try! The 90 nanometre technology refers to the thickness of the lanes we use, basically, to do the wiring on a silicon chip. To put that into perspective, in an area the size of your thumbnail, with Dothan, we can cram 140 million transistors into that die size. We are talking very small geometries.

The problem comes when you get these silicon atoms, if you link them in the manufacturing process they are basically crunched together and so an electronic signal basically struggles to get through those atoms, because they are all bunched together.

Think of it like going through a crowd of people at a shopping arcade. What this strained silicon technique does is to stretch out those atoms so that there is a lot more spacing between these atomic structures. And that way you can weave in and out, rather than barging through, and get from A to B much faster.

That's probably the best analogy I can do!

And ultimately this will impact on performance?

MacDonald: Oh yes, because when you are waiting for a signal to go from A to B it will get there nano seconds later than with strained silicon. Basically we see the benefit in two areas. In higher performance, and for a notebook user, because it is more efficient, it is getting the answer and going into idle and saving on battery.

The theory we have for notebook design is 'hurry up and get idle', 'huggi' we call it. Whatever you are trying to get done, do it is as quickly as you can, and go down to deep sleep mode. That means the notebook is available and not burning power.

Author: Alun Williams