Forgot to tell that vertical communication between stacked wafers is now
a
reality:
http://www.eetuk.com/tech/news/dev/OEG20030207S0034
Inge
-----Ursprungligt meddelande-----
Från: Ingemar Hernefjord [mailto:[log in to unmask]]
Skickat: den 2 december 2005 09:03
Till: 'TechNet E-Mail Forum'; [log in to unmask]
Ämne: SV: [TN] Moore's law and PWBs
Right Joe,
I participated myself in a project some years ago, the development of
something that reminds of a Neumann computer, but better. The genious
person that ran the project got support from our financial elite and it
was a
thrilling time. He hired programmers from India and signed a contract
with a semi company and a new language was created, the H language,
which
was said to be superiour to anything else. But he made a mistake, when
it
came to packaging and cooling and housing of the supercomputer, he
underestimated both time and cost. I told him we had to abandon the
traditional ways of thinking and do something quite new to take care
of the enormous network of I/Os and power dissipation, but he did not
listen. He began to run out of time and money and wanted a cheap
solution.
I proposed inventing vertical connections through the wafers, so they
could
speak vertically also, instead of making all connections to the perifery
with its long distances. I proposed cooling the wafers with a salt
solution,
and that we had to fill the gap between the wafers and the coolers with
Helium etc. The stack of several 12 inch wafers developed kilowatts when
working at maximum. But he did not want to spend a lot of money on
'simple mechanical things' and the result was that he ran out of the
deadline the investors had set to see result. The housing of the monster
I think got only some percentage of the total. The money flow was
stopped
and the whole project frozen for the future. And maybe this was a good
thing. If (I say if, because I don't know)this project had been a
success,
it had outnumbered Cray and other supercomputer manufacturers. Any
military
aircraft could have their own 'Cray', not bigger than a PC tower on your
desk! I spoke with the inventor and got permission to send a pic of part
of the superwafer to you. He said, doesn't matter, noone will
understand a
thing! And that's true, instead of the familiar processor pattern with
its
zillions of parallel lines and 90 degree turns, and transistor cells,
this
seems like a meaningless mess without structure. Likewise, I don't
understand anything either, I was not involved in those secrets, and my
competence was unsufficient anyhow, at that time being 'just a
packaging'
expert. My thick map filled with hundreds of pages with calculations and
proposals...I thought it could be a sweet memory, but finally I
destructed
everything. And so did most of the others...I have one wafer left.
Inge
-----Ursprungligt meddelande-----
Från: TechNet [mailto:[log in to unmask]] För Joe Fjelstad
Skickat: den 2 december 2005 00:16
Till: [log in to unmask]
Ämne: Re: [TN] Moore's law and PWBs
Along that line there is interesting article in Electronic Business this
month titled
"The Future That Never Was"...Seven products that could have changed the
industry but didn't by Geoffrey James
Somethings actually do happen but have morphed by time and evolution.
---------------------------------------------------
Technet Mail List provided as a service by IPC using LISTSERV 1.8e
To unsubscribe, send a message to [log in to unmask] with following text
in
the BODY (NOT the subject field): SIGNOFF Technet
To temporarily halt or (re-start) delivery of Technet send e-mail to
[log in to unmask]: SET Technet NOMAIL or (MAIL)
To receive ONE mailing per day of all the posts: send e-mail to
[log in to unmask]: SET Technet Digest
Search the archives of previous posts at:
http://listserv.ipc.org/archives
Please visit IPC web site
http://www.ipc.org/contentpage.asp?Pageid=4.3.16 for additional
information, or contact Keach Sasamori at [log in to unmask] or 847-615-7100
ext.2815
-----------------------------------------------------
---------------------------------------------------
Technet Mail List provided as a service by IPC using LISTSERV 1.8e
To unsubscribe, send a message to [log in to unmask] with following text in
the BODY (NOT the subject field): SIGNOFF Technet
To temporarily halt or (re-start) delivery of Technet send e-mail to [log in to unmask]: SET Technet NOMAIL or (MAIL)
To receive ONE mailing per day of all the posts: send e-mail to [log in to unmask]: SET Technet Digest
Search the archives of previous posts at: http://listserv.ipc.org/archives
Please visit IPC web site http://www.ipc.org/contentpage.asp?Pageid=4.3.16 for additional information, or contact Keach Sasamori at [log in to unmask] or 847-615-7100 ext.2815
-----------------------------------------------------
|
