[Beowulf] Re: ECC Memory and Job Failures (Huw Lynes)

[Robert G. Brown]

On Fri, 24 Apr 2009, John Hearns wrote:

> 2009/4/24 Robert G. Brown <rgb at phy.duke.edu>:
>>
>>
>> I don't think memory is all that unstable, especially down where I live.
>> In Denver, maybe.  I think you need a lot of RAM, for a long time, to
>> see a lot of radiation induced errors, or a source of high energy
>> particles.
>
> I thought more of a motherboard of RAM chips - which were flat in
> those days (mumble) years ago, and putting a radiation source directly
> above them. Poor man's silicon strip detector.

I vaguely remember an article long ago where somebody opened a RAM chip
and hooked it up so that they could play with the timing refresh.  RAM
is sensitive to light.  They wrote an array of ones while projecting an
image onto it, waited a suitable amount of time, and could read the
image out of the memory in 1's and 0's where the light hitting the array
discharged the caps.  A poor man's camera.

Hmmm, let's see.  Yeah, there is even a patent somebody filed for this
(GIYF).

So this might work if you had enough flux.

Hah!  Google IS your mighty friend!  Google up:

Radiation Dosimetry Using Three-Dimensional Optical Random Access
Memories

Hmm, people make cheap neutron detectors out of DRAM.

So I guess this would work, but I still think you need a pretty peppy
particle.  The article suggests 0.5 MeV or up.

    rgb

>
> _______________________________________________
> Beowulf mailing list, Beowulf at beowulf.org sponsored by Penguin Computing
> To change your subscription (digest mode or unsubscribe) visit http://www.beowulf.org/mailman/listinfo/beowulf
>

Robert G. Brown	                       http://www.phy.duke.edu/~rgb/
Duke University Dept. of Physics, Box 90305
Durham, N.C. 27708-0305
Phone: 1-919-660-2567  Fax: 919-660-2525     email:rgb at phy.duke.edu