Re: how can a bit be off in memory?

CBFalconer wrote:
The Natural Philosopher wrote:
Charles T. Smith wrote:

... snip ...
So the image on disk was fine, but the image as READ by the system
was corrupt, until you rebooted.

Actually, I guess I'll buy the arguments of the ECC (EMA?) posters
that it shows why ECC is a good thing.
I don't. I am fairly certain you had a corrupt transfer off the disk.

Doubtful. Disk extractions are checked (via a checksum/etc).

Up to the disk controller, yes. Then they are straight 8/16 bit transfers. Into maybe ECC memory.

The transfer is the weak point.

Without ECC memory errors go totally unnoticed, except for fouled
results.


Fouled transfers go unnoticed no matter what the RAM (as I discovered when it happened to me).
.

Relevant Pages

  • Re: Buffered and Registered Memory: Is it worth the expense?
    ... corrupt the disk and if it hits an OS files like the registry, ... My big machines have ECC and log everyting. ... Recoverable ECC hits *not* considered normal. ...
    (alt.comp.hardware.pc-homebuilt)
  • Re: how can a bit be off in memory?
    ... snip ... ... I am fairly certain you had a corrupt transfer off the disk. ... Disk extractions are checked. ... Into maybe ECC memory. ...
    (comp.os.linux.misc)
  • Re: how can a bit be off in memory?
    ... I am fairly certain you had a corrupt transfer off the disk. ... Into maybe ECC memory. ... Fouled transfers go unnoticed no matter what the RAM (as I ...
    (comp.os.linux.misc)
  • Re: Quad Questions
    ... I spent 3 years faffing and broke down eventually... ... > ECC versus Non ECC - does it matter? ... The air coming out of the ...
    (uk.comp.sys.mac)
  • Re: SUN Blade 100
    ... [SNIP] ... that explains the problem - I don't have any ECC on hand. ... I think there was a jumper on the mb to allow for unbuffered memory. ... Our OS who art in CPU, UNIX be thy name. ...
    (comp.sys.sun.hardware)