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.

Politics : Formerly About Advanced Micro Devices -- Ignore unavailable to you. Want to Upgrade?

To: Bill Jackson who wrote (69567)	8/23/1999 2:04:00 AM
From: Petz	Read Replies (3) \| Respond to of 1574061

Bill & Dan3, dropping geometry down by 28% (0.25 to 0.18) reduces capacitance down by 48% (proportional to size*2). Current in general has to drop linearly if both processes are aluminium. Thus, in general, max frequency should scale as the inverse of feature size. (Switching frequency is proportional to I/C.) But with copper, the power loss IIR through conductors is reduced by even more (because I and* R are smaller) and power loss due to charging the capacitance drops by the square of the feature size as well. Thus power consumption can drop my much more than a linear amount. This is especially beneficial for the Athlon, because I believe the thing limiting the MHz on 0.25æ Athlons is the power dissipation. PROOF: supercooled Athlons are running at 1 GHz right now. So why doesn't Intel think its worth it at 0.18æ? Because 1)power dissipation is not the limiting factor in Intel's sixth generation design as it is in the seventh generation Athlon, so the MHz gain on the Coppermine would only be half that of the Athlon. 2)Intel's potential gain of 10% in MHz didn't appear to be necessary to stay on top of the heap in 0.18æ; How wrong that logic will turn out to be! Petz