Silicon Investor (SI) -- The First Internet Community

STOCKTALK

We've detected that you're using an ad content blocking browser plug-in or feature. Ads provide a critical source of revenue to the continued operation of Silicon Investor. We ask that you disable ad blocking while on Silicon Investor in the best interests of our community. If you are not using an ad blocker but are still receiving this message, make sure your browser's tracking protection is set to the 'standard' level.

Technology Stocks : Advanced Micro Devices - Moderated (AMD) -- Ignore unavailable to you. Want to Upgrade?

To: Pravin Kamdar who wrote (228866)	3/25/2007 9:05:24 AM
From: combjelly	Respond to of 275872

"It's the other way around. Random variations due to "not enough atoms" is what requires statistical modeling" That isn't my assertion, it is Sarmad's. I haven't seen the papers in question. But, for both the reasons you assert and the implications I point out in a later post, I think he isn't ready the paper properly. It isn't like 65nm is the first node where process variation has introduced difficulties in the model. And, quite frankly, I find the idea that the variations can't be modeled to be absurd. But, again, I haven't read the papers in question.

To: Pravin Kamdar who wrote (228866)	3/25/2007 9:59:17 AM
From: Sarmad Y. Hermiz	Read Replies (2) \| Respond to of 275872

That was what I said. I'll try to re-state it. My understanding is that from one wafer to another, there are variations due to random atomic states. Because (at 65 nm and below) there are only a few atoms in each structure, the randon variation in a single atom become significant. So making adjustments (as in AMD's process control method) become less effective, and even self defeating. Because the next wafer does not reproduce the random states of the one where measurements were made.