[Gllug] server sleeping

[Nix]

On Thu, 14 Sep 2006, Richard Jones stipulated:
> On Thu, Sep 14, 2006 at 01:02:42AM +0100, Nix wrote:
>> Hell, memtest86 *and* memtest86+ *both* missed a faulty RAM chip that
>> was bad enough that md5sums of a 10Mb file came out different four times
>> out of five. It'll basically only find RAM that's *really* bad: an area
>> of RAM that only goes wrong if you write the wrong specific pattern to
>> it, or to neighbouring cells, is likely to be missed.
> 
> General memory testing is a hard problem.  Hardware can hit memory
> much faster than software.  (I used to design and implement such
> testing hardware).

... and many's the time I've wished I had such hardware, but it
*costs*. It's much cheaper just to replace the RAM :/

I'd venture to suggest that *only* hardware can reliably test
RAM chips; in fact in the limit only the chips themselves can
do it (ECC, and so on).

-- 
`In typical emacs fashion, it is both absurdly ornate and
 still not really what one wanted.' --- jdev
-- 
Gllug mailing list  -  Gllug at gllug.org.uk
http://lists.gllug.org.uk/mailman/listinfo/gllug