lkml.org 
[lkml]   [1996]   [Apr]   [23]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
/
From
Date
SubjectRe: memtest86, built into kernel
On 22 Apr 96 at 17:29, Matthew J Brown wrote:

> Matti E. Aarnio writes:
> >> No mainframes ran with ECC memory boards, so hardware faults got logged.
> >> PC's _had_ parity to show you errors in RAM. Now they don't bother (I guess
> >> because the people who make the PC's make so much bad memory they dont want
> >> you to find out).
> >
> > http://www-cs.intel.com/oem_developer/chip_pci.htm
> >
> > Pick "General" from that (the URL is VERY long..), and
> > go see about "Evaluating the Need for Parity in Desktop PCs"
>
> Basically they say there that with modern SIMMs the MTBF for a soft
> error (ie. alpha particle corruption) is between 12-30 years of
> continual system use.

Interestingly, this would require to test the RAM for at least 30
years to be sure. No company will guarantee the MTBF. BTW, as you
surely know the mean value alone does not say to much as long as you
don't know the variance. This is the time of business graphics...

>
> Given that it happens so rarely, that parity is only 50% likely to
> catch the error anyway, and that parity requires an extra 12.5% DRAM,
> it doesn't seem worth it to me. ECC is more useful, since it will
> correct single-bit errors rather than just hanging.
>
> -Matt
Ulrich


\
 
 \ /
  Last update: 2005-03-22 13:37    [W:0.035 / U:0.368 seconds]
©2003-2020 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site