On Tue, Jun 25, 2013 at 4:28 AM, Gene Heskett <ghesk...@wdtv.com> wrote:

> Any control computer in a nuke facility really should be rad hard, so that
> it can continue to function right up to within a millisecond of being a
> glow in the dark crater in the ground.  Those PDP-11's aren't.

I don't know of computers running a realtime control loop of a nuke
facility, but I've been running data acquisition in such ever since I
was a baby. I started on a PDP, then used a VAX and now it's
wall-to-wall Linux PCs. We do see single-bit upsets and account for
them, but the rate is measured in microHertz. If the radiation is low
enough so that you can sit a person there, you can sit a computer.
This by the way is a huge challenge for space travel: outside of LEO
the radiation levels are nasty and affect heavily both people and
electronics.

Funny enough, the old semiconductor technology had large cell areas
and charges, which gave it relative immunity to charge deposition
caused by ionization events. The newer technology shrunk the feature
size and unit charges, but  it turns out that smaller size makes it
harder to hit an individual cell, and also apparently there's more
ECC-style error correction in the data paths of the more modern
designs, so the bottom line seems to be that the radiation sensitivity
did not increase by much.

------------------------------------------------------------------------------
This SF.net email is sponsored by Windows:

Build for Windows Store.

http://p.sf.net/sfu/windows-dev2dev
_______________________________________________
Emc-users mailing list
Emc-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/emc-users

Reply via email to