Carl, <The entrance and exit pins for a Rambus channel signal are most certainly not the same signal pin.>
I think we have two different perspectives here. You are looking at the module socket, and yes, because of the two-ended nature of a RIMM, the signal pins have to be doubled from 37 to 74.
I, on the other hand, am looking at the whole system, where all I see are 37 continuous signal lines. Or, more specifically, 37 signals coming out of the memory controller. That's one reason why chipsets like the 840 have a very easy time grafting dual Rambus channels on the north bridge, because the pin count is so low. Try grafting dual DDR channels onto the north bridge. It's possible, if you don't mind the 200 extra pins or so.
In any case, we're both talking semantics here, because we can define and redefine what is meant by a signal pin. It's the stubless nature of RDRAM vs. the stubby nature of SDRAM that really muddies the waters here. In the end, it's a trade-off between the many factors of a system, as you have reminded me in the past. I just think your original post was somewhat subjective and artificial.
Tenchusatsu |
