lkml.org 
[lkml]   [1996]   [Nov]   [15]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
/
Date
From
SubjectRe: Linux & ECC memory

On Thu, 14 Nov 1996, Richard B. Johnson wrote:

> Since RAM access happens so quickly, keeping statistics on corrections
> that have occurred could seriously impact performance. The performance of
> VAXes was not seriously affected because they were soooooo sloooow to
> begin with.

I'm not so sure about that. Incrementing a counter on an ECC trigger
doesn't seem impossible. (Hmm. Unless the ECC isn't a compare-and-fix
cycle, but rather a discrete chunk of logic that always applies the math
without looking at what it is doing. I'm not knowledgable enough about the
algorithms involved to say for sure.)

And, to switch it around, if the ECC is being used often enough to degrade
performance, wouldn't you want to know about it? (Especially since an ECC
is by its nature a statistical operation. Multi-bit errors can go
undetected, and even be miscorrected.)

--
Kenneth Albanowski (kjahds@kjahds.com, CIS: 70705,126)



\
 
 \ /
  Last update: 2005-03-22 13:38    [from the cache]
©2003-2011 Jasper Spaans