Path: utzoo!utgpu!jarvis.csri.toronto.edu!mailrus!tut.cis.ohio-state.edu!bloom-beacon!gatech!hubcap!ncrcae!ncr-sd!hp-sdd!hplabs!hpfcdc!hpldola!hp-lsd!prisma!mo
From: mo@prisma
Newsgroups: comp.arch
Subject: Re: quest for breakthroughs (long)
Message-ID: <2700009@prisma>
Date: 24 Feb 89 15:55:00 GMT
References: <740@tetons>
Lines: 16

Going very very fast is very very hard if for no other reason than
time of flight delays across realizable circuit board material
and delays going on and off chips reduces the "1-foot per nanosecond"
speed-of-light-in-a-vacuum to more like "1-inch per nanosecond."
It ain't quite that bad, but it is a good number to plan with.
In theory, one *can* make multilayer circuit boards out of laminated
teflon, TFE, and gold conductors, but they would be astronomically
expensive.

This alone makes many parts of the machine more than one clock away,
and you'd be surprised at how badly that complicates things.

Fairly quickly one comes to understand that the speed of the parts
is NOT the driver of the machine clock speed.

	-Mike