Server 2003 on a P2L97-DS shows only one CPU

[ChinnoDog]

I am trying to install Server 2003 on my P2L97-DS with 2 PII/300's. The
computer is detected as MPS Multiprocessor but task manager and system info
show it has only 1 CPU. (I've never seen this before! If it knows it is
multiprocessor, why doesn't it see the CPU????) I tried updating the bios
from the last production bios, 1008, to the latest beta bios, 1010 beta 4.
System info then showed that this BIOS has ACPI support. So I started over
and did a fresh install. The same thing happens. I then attempted to
manually select the kernel by pressing F5 during setup. Setting it to ACPI
Multiprocessor causes setup to hang during the first part right after it is
finished loading drivers and is starting the kernel. (Meaning, before I get
any more prompts after selecting the kernel.) I tried changing the MPS 1.4
support off and on and tried turning plug and play off and on to see if that
would make a difference, but it doesn't. I reset all settings in my BIOS
and reconfigured in case there was a hidden setting left over from the
previous version of the BIOS. Nothing.

This PC runs Windows 2000 server just fine. I have been using this PC since
I built it in 1998 first as my workstation and now my server. When in 2000
it was using the MPS Multiprocessor kernel as well. (Except, of course, it
worked!)

Anyone have any idea how to fix this?

 

[ChinnoDog]

Yes, I am using Klamath processors. Well this blows major goats, especially
since WinXP ran just fine (SMP wise) and this is supposed to be the server
version. I considered getting two Celeron 550's with slotkets at one point
but chose to pick up the second PII instead. I guess I'm stuck on 2000
server forever. Or Redhat... but I already have one of those. Thanks for
the info.

 

[P2B]

Good to know that. So I'll probably be going straight to a P2B-D(S) rev.
1.06, skipping the P2L97-DS ACPI fix part.

I have several modified and tested for Tualatin P2B-DS 1.06 boards in
stock, PM me if interested.

As for the second processor
not being recognised, it took me a while to remember the probable cause:
You're using old Klamath core PIIs, right? These (along with PPros, PII
Overdrives and early Xeons) contain an MP related erratum for which the
workaround is no longer included with Server 2003. The OS works fine
with two Mendocino Celerons on a P2L97-DS here, MPS HAL too. (That's
what I call irony - Once cheapish Celerons that never officially
supported SMP run just fine in MP mode with Server 2003, while the once
high-end early Xeons don't.)

Ironic indeed... I wonder if it could be fixed with a hack?

 

[P2B]

Yes, I am using Klamath processors. Well this blows major goats, especially
since WinXP ran just fine (SMP wise) and this is supposed to be the server
version. I considered getting two Celeron 550's with slotkets at one point
but chose to pick up the second PII instead. I guess I'm stuck on 2000
server forever. Or Redhat... but I already have one of those. Thanks for
the info.

I still have my original modded-for-SMP Celeron 300As, and I doubt I'll
ever use them again. PM me if interested.

 

[Ixnei]

It's because Micros~1 pulls this sort of shit, over and over ad nauseum,
that I will NEVER purchase one of their OS's again (nor will I use one,
unless absolutely necessary). These sorts of actions are classic. Same
goes for Intel, not ensuring support for Xeon MP's in newer server OS's...

Yes, I am using Klamath processors. Well this blows major goats,
especially since WinXP ran just fine (SMP wise) and this is supposed to
be the server version. I considered getting two Celeron 550's with
slotkets at one point but chose to pick up the second PII instead. I
guess I'm stuck on 2000 server forever. Or Redhat... but I already have
one of those. Thanks for the info.

--
We HAVE been at war with Iraq for 13 years now, bombing their
country on at least a weekly basis.
"U.S.-led sanctions have killed over a million Iraqi citizens,
according to UN studies" - James Jennings
3,000+ innocent Iraqi civilian casualties can't be "wrong"...

 

[ChinnoDog]

I wish there was a hack. I tried a couple unlikely hacks myself... like
trying to put the Win2k kernel into Win2k3. This thing isn't worth
upgrading. Its only 300Mhz and 100% SCSI. Thanks for the offer though.
Maybe support will appear in the next release of Windows. After, all, the
ability to name folders in Win95 with alt+255 characters was fixed in 98 and
then reappeared in ME. (If you don't know that one, try it. Do it from a
DOS prompt and then try to delete the folder in Explorer.)

 

[ChinnoDog]

Not to be ungrateful, but its already slow enough as it is! Plus, without
L2 cache and only a 66Mhz bus between them I would think they would spend
their days waiting for data off the bus instead of executing opcodes.

 

[P2B]

Not to be ungrateful, but its already slow enough as it is! Plus, without
L2 cache and only a 66Mhz bus between them I would think they would spend
their days waiting for data off the bus instead of executing opcodes.

You're correct - your P2L97-DS would not be able to take advantage of
these processors. They are the 'A' version with 128KB of full-speed L2
cache (vs. 512KB half-speed on your PII 300s), and spent their days at
504Mhz (112Mhz FSB). Mind you, Server 2003 *would* be able to use them
both...

 

[P2B]

I wish there was a hack. I tried a couple unlikely hacks myself... like
trying to put the Win2k kernel into Win2k3. This thing isn't worth
upgrading.

I tend to agree - with a maximum FSB of 83Mhz, your upgrade options are
severely limited.