I have just acquired a pair of Slot B modules marked 833Mhz. / 4MB cache.
These were purchased as an upgrade for my API Networks UP2000+ motherboard.
The motherboard is known operational (at least with my current 667Mhz. 2M
cache modules) and has firmware A5.8-81 installed.
The new modules do not appear operational and the seller will accept a return.
However, before I do so can anyone suggest any steps to better diagnose the
situation? With the original CPUs, I can see the SROM leds flicker for a
fraction of a second, then go out just prior to the BIOS signon screen. With
the new CPUs, the leds flicker and stay illuminated. The system never
proceeds past that point.
Any advice appreciated!
Steve
Did you set J31 accordingly? It needs to be closed for the 833MHz
modules.
From what I understand, there are some differences between the UP2000
and the UP2000+. They are not the same. For starters:
The UP2000+ has a higher clock speed
Fewer 64bit PCI slots
Now keep in mind they make the UP2000+ not only with the 833mhz cpu
but also the 667 and 750... somewhat confusing...
*Despite the claims, both had onboard SCSI and used the same memory,
etc.
Not sure I follow your point. I have a UP2000+, which definitely claims to
support the 833Mhz. CPU with up to 8MB of cache.
Steve
Yes, definitely. The documentation associates the need for this jumper with
installation of a 21264B cpu, not with the speed per se. However, a printed
document with the CPU stated that J31 must be shorted. I just take in on
faith that this is a "B" model as opposed to a 21264A. The module does not
indicate this one way or the other.
Steve
According to Wikipedia, the 833 MHz CPUs are 21264B definitely. Does
this change anything?
>>> Did you set J31 accordingly? It needs to be closed for the 833MHz
>>> modules.
>> Yes, definitely. The documentation associates the need for this jumper with
>> installation of a 21264B cpu, not with the speed per se. However, a printed
>> document with the CPU stated that J31 must be shorted. I just take in on
>> faith that this is a "B" model as opposed to a 21264A. The module does not
>> indicate this one way or the other.
>
> According to Wikipedia, the 833 MHz CPUs are 21264B definitely. Does
> this change anything?
Just confirms that I had the jumper set correctly!
Steve
Have you tried simply switching the cartridges around? If you have a
'flaky' cpu cartridge, often switching them around will allow it to
boot. I had a similar issue with my UP2000. CPU1 was buggy as hell and
the box wouldn't boot at all if I moved it to the CPU0 position. CPU1
did eventually burn out... then several months later CPU0 did
likewise :-(
> Have you tried simply switching the cartridges around? If you have a
> 'flaky' cpu cartridge, often switching them around will allow it to
> boot. I had a similar issue with my UP2000. CPU1 was buggy as hell and
> the box wouldn't boot at all if I moved it to the CPU0 position. CPU1
> did eventually burn out... then several months later CPU0 did
> likewise :-(
Yes, that was one of the first things I tried. After discovering that it
wouldn't start with both of them, I tried them in CPU0 one at a time. No joy.
I acts like the SROM load (microcode, I assume) is hanging. With operational
processors, the SROM status LED on the motherboard goes out very shortly after
reset. With either/both of these, it stays on.
One final mystery: The newest firmware I could find for the board was
A5.8-81. The documentation with the CPU mentions A5.8-332.1-61. I can find
no reference to that number anywhere. I don't understand the numbering scheme
and perhaps it was an older version (-61)?
Steve
I read that as well. I can only assume that 5.8-81 is newer and
therefore the SRM/SROM isn't the problem since I know that people have
used 833MHz CPUs in the UP2000+ and cannot find any mention of
5.8-332.1-61.
>
> Steve
It has (4) 64-bit PCI, (2) 32-bit PCI and one lonely ISA 16-bit slot at the
end (either/or with one of the 32-bit PCIs).
Why do you ask? Just curious.
I think he's asking to confirm that you have a UP2000+, as the UP2000
and UP2000+ have different numbers of 64-bit PCI slots.
Yes, if that were the case it would certainly account for why the CPUs weren't
operational. The board is definitely a UP2000+, with 'Rev 0B' on a sticker at
the rear.