X-Message-Number: 28598
From: 
Date: Thu, 26 Oct 2006 14:39:22 EDT
Subject:   Hypercomputing

Hypercomputing.

I think cryonet readers are faliliar  with the basic idea of quantum 

computing : With entangled states of 1's and 0's  an N bit circuit can perform 2
power 
N simultaneous computations. We have no  such computers mainly because 

quantum states are extremly frail and can't be  stabilized long enough to to 
complex 
operations on them.

In the  sixties, Lockheed had produced another kind of parallel computer. it 
used light  in waveguides. Depending on the pipe diameter, ther may be one or 
more  polarizations in the guide. If the diameter is half the wavelength, 

there is one  polarization and one information stream. Larger diameter allow for
more  polarizations, each with its own information content. The polarization 
number,  expand (from memory) as 2 power the guide diameter counted in half 

wavelength.  Assume near infra-red ligth at one micrometer is used, a channel 
five 
micrometer  in diameter would carry 2**10 = 1024 polarizations. With ten 

micrometer, the  number jump to one million and so on. The Lockheed prototype 
was 
able to carry  and compute on 2000 polarizations at the same time.

The problem  with such a device is at the entry and the end. each 

polarization must be  introduced and then extracted, this needs a very large 
electronics 
system, this  one negate any benefit in the optical computing. Add to that 

that unil recently,  not all operations could be performed in the optical 
domain, 
asking for more  mixer and separator steps. Now, this problem seems solved 
that let only the  entry and exit difficulties. Why not use the quantum system 
here ? Mixing and  unmixing are simple and fast operations, a quantum system 
would not hit its time  limit on that task.

An optical computer with 30 micrometers  waveguie would process one billion 
of billion operations simultaneously, a 60  bits quantum system would feed it 
and extract the results. This sems well in the  actual technological 

possibilities. Even if the optical part run at only one  million operation per 
second on 
each channel, the processing power would be in  the 10**24 operations per 
second. This would be interesting for both : Uploading  and full body repair 
control.

Yvan Bozzonetti.  




 Content-Type: text/html; charset="US-ASCII"

[ AUTOMATICALLY SKIPPING HTML ENCODING! ] 

Rate This Message: http://www.cryonet.org/cgi-bin/rate.cgi?msg=28598