Although this doesn't belong to NetBSD that much anymore, here I am
once again... I got MemCheck12.lha from Aminet and it reports errors
from various addresses (on motherboard memory), but sometimes it
does _not_ report errors at all.

Any good guesses which might be the cause?  Now that it doesn't report
errors every time and on same place, it cannot be a totally blown chip
(or perhaps memory chips work in a strange way).  Could it be just
overheating, Warp Engine (esp. 040) runs quite hot?

And how is the memory organized on the motherboard (A3000)?  Is the
memory on bank 0 in 0x07000000-0x073fffff and inside a bank, does
it go linearly?

Quite often the faulty address is 0x0704???? which might be the chip
under Warp Engine (if I guessed the memory organization right) and that
is one hot place to be... sometimes the address is 0x078?????, 0x07a?????
and 0x07e????? (so, again if I guessed right, the bank 1 from 0x07400000
to 0x077fffff is the only one without problems).  But I tested this only
about 15-20 times, so perhaps there are problems in every bank.

Anyway, I'd be glad to hear suggestions on what to do... Buy a new Ramsey?
Try again to pull&push every zip in its socket?  Something else?


But I must say that I'm starting to get enough with this problem
(it's been here for months, first I blamed SCSI cable because I thought
NetBSD reads data badly and that's the reason for every crash), so if there
is no simple solution I'll go and buy second 16MB SIMM (it would of course
be much faster than motherboard ram).


	Mika Kortelainen