Re: Memory and socket problems.

Steve O'Neill (soneill@netaxs.com)
Sat, 23 Nov 1996 08:28:42 -0500 (EST)

On Fri, 22 Nov 1996, James M. Darlack wrote:

>
> Hi Linuxoids!
>
> My DEC Alpha 166 MHz UDB refuses to recognise 2 32 mega byte true
> parity RAM chips.
>
> The 32 MB ram chips were installed with 2 4 meg chips, or with 2 8 meg
> chips. In each case the UDB thinks that the 32 meg chips are 8 megs.

Interesting. I've had the same problem with 16 Meg SIMMs. When I installed
them with the original 8 Meg SIMMs in place, the system saw a total of 32
Megs, not 48 Megs. If I put the original 4 Meg SIMMs in, the system saw a
total of 20 Megs. Swapping the 16 Meg SIMMs position around made no
difference. I have no indication from the system that there's anything wrong
with the 16 Megs, just that it won't see all of their memory.

>
> The amount of memory the UDB recognizes during boot time, via the
> Supplementary Menu, under Hardware Configuration.

I tried telling ARC how much memory I had, in the hope that it would get
milo to see that much, but, no go.

>
> Has anyone seen this problem, or how to cure it?
>
> I called the company where I purchased the Alpha, and they know of no
> jumper of any other option that must be set for large memory chips, or any
> other thing to be done.

I have a suspicion that the problem may lie in one of two areas: either in
the way the Alpha lays out its addressing, or in the physical addressing on
the SIMMs themselves. In the first case, it may simply be a matter of
configuring the firmware to "see" the additonal memory. The trick is
figuring what to set. If the second is true, then that would mean that just
any ol' true-parity SIMM won't work in the UDB. Since others have
successfully upgraded their memory without mentioning that there was
anything special about the SIMMs they used, I doubt if this is the problem.

Or, it could be none of the above. I haven't had any time to look further
into it, so I still don't know why the machine won't see all of the memory.
When I get a chance, I want to try the SIMMs on some other box, possibly a
PC, just to assure myself that they really _are_ ok. Have you been able to
verify that your SIMMs are good?

...(stuff deleted)

Steve O'Neill

--
To unsubscribe: send e-mail to axp-list-request@redhat.com with
'unsubscribe' as the subject.  Do not send it to axp-list@redhat.com



Feedback | Store | News | Support | Product Errata | About Us | Linux Info | Search | JumpWords
No Frames | Show Frames

Copyright © 1995-1997 Red Hat Software. Legal notices