Carl Lowenstein wrote:
Reality check. I haven't seen a computer in which the processor and memory were 6 to 12 inches apart (15 to 30 cm) since late PDP11 days (mid 1980's).
That last DIMM socket (out of 4) is about 4 inches from the processor pins (which are *not* directly under the edge of the heatsink). I just took a ruler to my AMD board. I haven't seen many motherboards where that isn't the case. You need a finite amount of space for the socket and for the airflow to dissipate the heat from the memory. That doesn't include any routing on the DIMM itself to balance the delays.
I'll take a quick check at Fry's next time i go through, but I still stand by my estimate that memory has a low 10's of centimeters route from the processor to the memory on commodity boards.
Another reality check. The speed of electromagnetic wave propagation in copper is considerably faster than the speed of the electrons.
True, but we're not just playing with waves and antennas. The proof of that is that you can stick a circuit at one of the wavelength nulls and the circuit still works. That doesn't happen when you are relying completely on waves.
In reality, the electron response time is dominated by wave propagation, but charging the load still requires electron drift velocities. This comes in around 1V/ns or so. This delay is *on top of* the wave propagation.
For a fairly standard board, propagation velocity is about 1/3 the speed of light in a vacuum.
You do realize, though, that you're quibbling with me over factors of 2 when I'm making engineering estimates that are, at best, accurate only to about a factor of 10?
-a -- [email protected] http://www.kernel-panic.org/cgi-bin/mailman/listinfo/kplug-lpsg
