Re: memtest86, built into kernel

Oleg Zabluda (zabluda@math.psu.edu)
23 Apr 1996 04:36:54 GMT

Matti E Aarnio (mea@mea.utu.fi) wrote:
: 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"
:

I've read this. It's full of bull. Basically Intel claims that
the DRAM error rate due to alpha particles for 16 Meg system
is on the order of 20 years, therefore parity checking is not needed.
Maybe at Intel memory errors are caused by alpha particles. But
on my desk they are caused by oxidation of tin SIMM sockets.
I know, I know, I should have bought tin plated SIMMS, instead of gold
plated ones. But with the price of memory at that time I went for the
cheapest ones, but with parity, which is needed anyway.
Even if a given system has a memory error every 20 years it means
that out of 7300 systems one has a memory error every day.

Oleg.