[Gllug] server sleeping

[Nix]

On Thu, 14 Sep 2006, t. clarke mused:
> Making a machine work hard with GCC or whatever may well cause a machine to
> fall over whereas memtest86 will not find a problem,  but that still, I guess,
> gives you the problem of identifying the faulty component!  Could just as well
> be cache as primamry memory!!

Indeed, although bad cache will cause failures *much* faster than bad
RAM unless the RAM is very, very bad.

(Can memtest86 differentiate these cases? A lot of machines let you turn
off L2 cache in the BIOS, but L1 cache... I guess if that's bad the
machine probably won't start at all, much as if the TLB or register file
is bad.)

>                                Also you cant guarantee that a heavy-duty
> compile will use ALL memory - the kernel for a start will grab a fair chunk that
> GCC will never see.

Indeed: it won't spot problems at the very top of physical memory, for
instance.

>                     So, IMHO, memtest86 is a useful tool even if not 'perfect'.

Oh yes.

-- 
`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