3 Nov
2005
3 Nov
'05
11:39 p.m.
Subject: Re: FPGA VAX update, now DIY TTL computers
From: "Chuck Guzis" <cclist at sydex.com>
Date: Thu, 03 Nov 2005 15:22:03 -0800
To: cctalk at classiccmp.org
On 11/3/2005 at 7:53 PM ard at p850ug1.demon.co.uk wrote:
If you ask me the real horror of parity memory is it only tells you disaster
happend just as your about to crash. It offers little to system reliability
and data integrity that real ECC can provide. Like Tim O' said years ago
it's all a plot to sell more ram.
Allison
There was a hrrible thing called fake parity
memory. It stored 8 bits and
calaculted the parity bit on a read (ignored the incoming parity bit on a
write) to keep the memory controller happy. Of course you can't store 9
bits in that type of device.
Of course you can, as long as the bits you're storing agree with the parity
of the 8 bits...
:) 
4 Nov
4 Nov
12:16 a.m.
Allison wrote:
If you ask me the real horror of parity memory is it
only tells you
disaster
happend just as your about to crash. It offers little to system reliability
and data integrity that real ECC can provide. Like Tim O' said years ago
it's all a plot to sell more ram.
Allison
Parity depends on if you would prefer a crash to silently generating
incorrect results.
ECC tries to hide the errors, but can still cause a crash if too many
bits change.
12:21 a.m.
On 11/3/2005 at 5:16 PM Kevin Handy wrote:
Parity depends on if you would prefer a crash to
silently generating
incorrect results.
(Insert favorite (perhaps apocryphal) Seymour Cray quote here)
ECC tries to hide the errors, but can still cause a
crash if too many
bits change.
We used to call it SECDED for exactly that reason. But if you're suddenly
encountering multiple bit errors where you had no errors previously, you've
got bigger problems.
Cheers,
Chuck
3:09 p.m.
>>>> "Kevin" == Kevin Handy
<kth at srv.net> writes:
Kevin> Allison wrote:
> If you ask me the real horror of parity memory is
it only tells
> you disaster
> happend just as your about to crash. ...
Kevin> Parity depends on if you would prefer a crash to silently
Kevin> generating incorrect results.
Kevin> ECC tries to hide the errors, but can still cause a crash if
Kevin> too many bits change.
Correct. And all error checking codes will fail to detect certain
errors.
Which is the right answer depends on what kind of errors you consider
likely enough to worry about, and what recovery you require.
paul
7091
days inactive
7092
days old
3 comments
4 participants
participants (4)
-
ajp166＠bellatlantic.net -
cclist＠sydex.com -
kth＠srv.net -
pkoning＠equallogic.com