very high GHZ cpus
john dobbs
GHZ speeds like a 16 Ghz or 30 Ghz desktop. I can only see optical
being the main method of signalling vs using electrical methods PC
motherboards with fiberoptics replacing the copper for data
communications would seem like the only way out. and what about
cooling, will desktops now ship with flurocarbon cooling systems. It
almosts looks like a step backwards PCs getting bigger,heavier. I
wonder how fast a spellcheck will take :)
Del Cecchi
The 30 GHz will only be inside the chip. The ISA bus will continue to run at
4.77 MHz.
--
Del Cecchi
cec...@us.ibm.com
Personal Opinions Only
Rupert Pigott
news:apomre$181a$1...@news.rchland.ibm.com...
Mind if you send you the medical bill for repairing
the damage I sustained falling off my chair ? :)
Cheers,
Rupert
Ketil Malde
> In article <9e63d1b0.0210...@posting.google.com>,
> jdobb...@yahoo.com (john dobbs) writes:
> |> how will intel and motherboard companies deal with CPUS reaching high
> |> GHZ speeds like a 16 Ghz or 30 Ghz desktop. I can only see optical
> |> being the main method of signalling vs using electrical methods PC
> |> motherboards with fiberoptics replacing the copper for data
> |> communications would seem like the only way out. and what about
> |> cooling, will desktops now ship with flurocarbon cooling systems. It
> |> almosts looks like a step backwards PCs getting bigger,heavier. I
> |> wonder how fast a spellcheck will take :)
I don't think cooling will be much of a problem with optical systems:
instead of copper or aluminum heat sinks to lead away excess
electrical energy, one just needs a transparent case to let out the
excess light. (And perhaps a clever system of mirrors, to lead it out
of the building)
With the advent of nano technology, one could also utilize mechanical
signals. While even light takes some time to travel, if you push one
end of a solid rod, the other end will move instantaneously! The
trick is to make the rod light enough so that inertia won't stop you
from getting high data rates. I think there's a Russian company that
has some patents on this technology.
> The 30 GHz will only be inside the chip. The ISA bus will continue to run at
> 4.77 MHz.
But surely fiber would allow much larger (distributed, say) ISA buses?
I think mechanics make for a better bus interconnect, and you can of
course have electrical signals in the rods for backwards
compatibility, and for high-bandwidth/high-latency communication.
-kzm
--
If I haven't seen further, it is by standing in the footprints of giants
Bernd Paysan
> While even light takes some time to travel, if you push one
> end of a solid rod, the other end will move instantaneously!
Nice try. Solid rods propagate pushes at sonic speed, and even if you had an
infinitive hard material, the limit for sonic speed would be vacuum light
speed. Sonic speed in usual materials is about 5 orders of magnitude slower
than light.
The only "instantaneously" (non-local) transmitted stuff I know of is the
EPR-style Quantum Mechanics measurement on entangled particles. The
measurement of one particle influences the other "instantaneously", but
unfortunately, you can't know the original state, and the only way to find
out that another measurement has happend is to correlate the results.
It's also dubious that the world really works as QM predicts. There are only
two sufficiently successfully experiments to this problem (dynamic EPR -
static EPR doesn't have a real problem). The first one (Aspect et al) had
6m distance with 50MHz periodic switching - and was quite fast debunked by
Anton Zeilinger. A student of him then made another experiments (a few
years ago) with a random switch - something that should be fool-proof.
I still have a few doubts with that experiment (there are some dubious parts
used in the photon path, and the photon visibility is too low; it could be
as well a QM interference experiment instead of a EPR experiment), and
especially wonder why Zeilinger and his team haven't just tried to repeat
Aspect's experiment with other l/f parameters *first*. The 6m/50MHz
correlation looks a lot like "we tuned this and that until we got the
result we were looking for".
Experimental physics is a lot like engineering. It's about achieving a
result. You can either make simple experiments that have obviously no flaw,
or complicated experiments that have no obvious flaw. At least I don't feel
very comfortable with a world that has ghost-like non-local effects.
--
Bernd Paysan
"If you want it done right, you have to do it yourself"
http://www.jwdt.com/~paysan/
Bill Todd
"Ketil Malde" <ket...@ii.uib.no> wrote in message
news:eg7kg0u...@sefirot.ii.uib.no...
By the time these frequencies are reached, object technology will have
become passe. ISA buses will thus be irrelevant, having been replaced with
ISNOTA buses (i.e., switches).
- bill