JC, "...if AMD can produce 70-80nm gates, why can't Intel if Intel's process is equal or better? - Elmer"
This is a big fallacy, and the major point on
contention between Intel "process technology"
zealots and AMD admirers.
Intel zealots can't stand that FACT that AMD
process technology is better, and _ALWAYS_ was,
although it may sound paradoxical.
The simple truth is that the process advantage was
the only tool where AMD could compete with Intel
in the copy-CPU era of 286/386/486. And they did it
successfully.
In K6 era,
when the 5-stage pipe of K6 had serious drawbacks
against 10/12 stage pipe of the PPro/II/III, AMD
managed almost to reach frequency parity. Given
the "8-level of logic rule" for PPro pipes versus
20-25 levels of logic in K6, AMD advantages in
transistor/layout/interconnect technology was
as high as 2.5X of the Intel. I guess some
process technologists at Intel and their
"consultants" were not comfortable with this
fact.
The issue of (transistors) # (pipe length) was
greatly discussed on AMD thread in K6-era, but
few FAB technicians (with PhDs) did not get the
idea despite of elaborate attempts of the thread
to educate them.
With the pipe length parity between P-III and Athlon,
the AMD process technology advantage (40+%) was clear
for anyone.
Now, when the P4 pipe is twice as long, these morons
are back. These guys have an agenda multiplied by
deep ignorance, so it is totally useless to argue
with them, only curse them away.
Regards,
- Ali |
