Pentium Pro ES CPUID's

[gg1978]

I've been doing some experiments to get the CPUID's of my ES PPro's, or the older ones at least.

Q0720 is 0610
Q0755 is 0610
Q0815 is 0611

It was interesting, as the 0610 PPro's wouldn't boot a 32 bit linux distro, while the 0611 or later would boot OK with no problem at all. So i had to haul out an old hard drive with Win ME on it, and got the CPUID's that way.

I also have a Q0766 that will need its pins massaged before i will be able to get it into the CPU socket and see if it is a 0610 or 0611. I've got a nice collection of PPro motherboards to choose from, for this i've been using an Iwill P6AN. This Q0766 is also marked A4.1, which i presume is the revision, as B0 was the initial released Pentium Pro's.

Does anyone know of any other tools than CHKCPU that will work in either real mode DOS or in Windows that will get more information ??

[debs3759]

Have you tried CPU-Z? You might need an older version.

[Neon]

maybe ASTRA

http://sysinfolab.com/

[Neon_WA]

[gg1978]

The Iwill works good, except for the Linux reboot issue.. I've gotten a bit more into what causes it, apparantly when Linux doesn't recognise your CPU correctly it tends to crash like that.. It shows up as a "Pentium Pro stepping 00" at boot..

I also have an Intel Performance/AU that works good with these ES's, it doesn't reboot when the Linux crash happens, so you can actually read the text of what it was doing last when it died..


Yup, both the P6AN and Performance/AU have the VRM override capability, the P6AN allows setting any VID, while the AU just allows setting to 3.1.

The P6AN can run down to 50x2, which results in a rather slow 100Mhz PPro :p.. The Performance/AU allows 60x2 so far, it might have an undocumented 50MHz setting as well.


I'd guess A4, based on the ink marker writing on the Q0766 being A4.1, which i guess would be a minor spin on the A4 stepping.

The Q0720 is quite a bit older, from week 11 of 95, while the Q0755 is from week 26 or so. I don't have them in front of me so going by memory here..

[gg1978]

On a side note, i wonder if the totally unmarked PPro ES's are even older than Q0720?? I know i've seen a couple on the web. It would be cool to see a 0601 CPUID or something like that, for an A0 PPro if such a beast exists.

[Neon_WA]

[Antinomy]

I think you (we) should start thinking of a database for such stuff.
Each one of us has some info that's not what you can google within 10 minutes. And it might be worth collecting.

[gg1978]

well, check out the two latest submissions to the CPUID database

I should have noted earlier that these chips didn't like booting into Windows ME regularly, i.e. they prefer safe mode.. I did get the Q0755 to get into Windows at 120MHz, 3.1V, though it was quite unstable at those settings, though it worked to submit the CPUID to cpu-world at that speed, though 100MHz would be more realistic for stability.. As for the Q0720, it would only run at 100MHz 3.3V, and even then it was barely stable enough to submit the CPUID info. I'm starting to think the Q0720 is a different stepping, if it's barely workable at 100MHz..

But it's there if you want to have a look at it

[Neon_WA]

nice one mate

I notice that Q0720 & Q0755 come up with one extra feature to production Pentium Pros in CPUid database
Advanced programmable interrupt controller

Not sure if due to being early or a feature that is only flagged in ES

you may find this spreadsheet useful
https://docs.google.com/spreadsheets/d/1LIIMep5IqPZKPpgWTt7SmJnZUf2R3ZJnhgPbAIxreRc/edit?usp=sharing

[gg1978]

Perhaps the APIC on these ES PPro's is busted, and that's why it's disabled on the production CPUs ?? I'll try some of my later ES's and see if they have the APIC feature or not..

[Antinomy]

I was googling for an PPro ES Sspec and got over here: http://www.intel.com/design/archives/processors/pro/qit/index.htm


http://www.intel.com/design/archives/processors/pro/qit/database/Q0815.htm

I think I'll grab the whole part of the site.