Talk:Pumping (computer systems)

Reasoning behind quad pumping

I find it difficult to understand why quad pumping is ever performed. Double pumping makes a lot of sense; the data and clock wires both have the same maximum frequency. With quad pumping, all the data wires have to be able to handle twice the frequency of the clock in the worst case, and it seems to me that any sensible implementation would involve synthesising a double frequency clock. So why not just double the clock and use double pumping? Is this marketing, or is there a good technical reason? Finally, don't mother boards with 1066 MHz FSB Intel processors have a lot of trouble with ringing and RF (very high frequency radio frequency) effects when the data signals will be effectively 533MHz square waves? Is this possible because they are only needed for the very short distance between the processor and the northbridge? (Some of this would obviously belong on the Front Side Bus page.) --Mike Van Emmerik 08:27, 18 October 2006 (UTC)