outram@cernvax.cern.ch (Nicholas Outram) (05/07/90)
Hello there,
I have a piece of information that may be of interest to
this newsgroup so I am posting it. It was published in a magazine a friend
bought home last night. I think it was called 'physics today' - sorry
for not having a more specific references, if anyone is interested I will get
full details.
Info.
-----
There where two articles of interest, both about the use of Josephson
Junctions in the production of components for computers. Basically there
has been a Japanese breakthrough in producing a 4bit cpu (bit-slice) and
a 1Kbit RAM. They are hoping to make a complete computer in a couple of
years.
The interesting point was the speed of these devices. They run at a clock
rate of 770MHz for the cpu and the memory has an access time of 500pSecs.
They have also managed to get a logic gate working at 1.5pSec for a reduced
dimension junction. The chips consume milliwatts of power, the disadvantage
being they have to be cooled to superconducting temperatures.
> The implications of this are enormouse...
1. IBM stopped its Josephson junction research in 1983 after having problems
with the lead alloy junctions. The Japanese are useing Niobium alloys.
2. The chips run 20 times faster than silicon and consume 1/2000 of the power.
3. All this development and breakthrough has been achieved with minimum inv-
estment. When the 'big-boys' catch on the rate of advancement will be
great.
This could be the point where silicon and even gallium arsenide bow out !
Imagine a parallel supercomputer running at 100GHz. Perhaps there is someone
out there who knows more about refrigeration technology, is it feasable to
shrink the required cooling down to desk-top proportions ?
Nick Outram.
--------------------------------------------------------------------------