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: fastpathguru who wrote (210918)	9/14/2006 3:53:39 PM
From: ChrisBBo	Read Replies (2) \| Respond to of 275872

I would intuitively expect frequency to scale with the length of the wires, not their area... For some reason, for each new process generation, it's been the aim to halve the area of all structures on the die. That's why we have the 500,350,250,180,130,90,65nm and future 45 and 32nm process generations. Notice how the ratio between successive generations is approximately the square root of 2. The result is a halving of area. Halving the area of a wire, halves its capacitance. If the wire is reduced equally (square root of 2 ~= 1.414) in both length and width, resistance stays the same. The combined effect is a wire that's "twice as fast". Another way of reducing capacitance is by using low-k dielectrics as insulators around the wires (do not confuse with high-k dielectrics used for transistor gates). A way of reducing resistance is to make the wires from better conductors - like cobber instead of aluminum. Just remember: area = capacitance = bad Length / width = resistance = bad The other part of the circuit equation is transistor performance. The amount of resistance and capacitance in the transistor. This is very simplified. For instance, the output resistance of a transistor isn't fixed. It varies non-linearly with voltage and current. The "drive strength" of a transistor, is the amount of current it can supply without its resistance rising sharply to intolerable levels. Raising voltage moves the point where this happens - effectively raising drive strength. But raising voltage increases leakage and power consumption in general. This can in turn be alleviated by improving the transistor with things like metal gates and high-k gate dielectrics. Another way of improving drive strength, is by improving the channel part of the gate. That's where strain, multi-gate, "notched" gate transistors come in.