GIGABYTE MZ31-AR0 REV 1.1 with Epyc Rome / Milan
- Thread starter Arcanum
- Start date
-
- Tags
- bios epyc motherboard
Notice: Page may contain affiliate links for which we may earn a small commission through services like Amazon Affiliates or Skimlinks.
I can guarantee you it is not AI. I said the sticker is official in the sense that it is a « foam » industrially printed label, nobody except a factory would produce such a label. This is probably 99% original from the factory. I am not home before Monday and that website only let you upload super low resolution picture, so I upload what I had available on my phone. Hi-res pics: Temp Image — image — IMG_1301.jpeg
and Temp Image — image — IMG_1156.jpeg (link valid 24h)
I can even take a video if you want. I have no reason to spread false information and I even offered to share dumps if needed…
and Temp Image — image — IMG_1156.jpeg (link valid 24h)
I can even take a video if you want. I have no reason to spread false information and I even offered to share dumps if needed…
there are no Gigabyte MZ31-AR0 1.1 with factory Rome support.
only 2.x and 3.x
Sticker is from Seller/dealer not gigabyte.
from a factory but not gigabyte.
and at the under side ? best joke ever.
i do not doubt it can run Rome.
but again; there are no Gigabyte MZ31-AR0 1.1 with factory Rome support.
edit: all MZ31 with rome support have AMI-MegaRAC license.
I would tend to disagree on that one, the font used on that sticker seems to be the Gigabyte font, which is not a public font. Faking it seems a huge effort for very little added value.
As far as I now, there are very few MZ31-AR0-00 with confirmed unlock, most of the confirmed unlock were HZ versions. It may be a very late series 1.1 before they switched to 2.x. I will check the serial next to ATX port when I go back home.
Anyway, I was just willing to help and could have done a VRM dump if it helped.
As far as I now, there are very few MZ31-AR0-00 with confirmed unlock, most of the confirmed unlock were HZ versions. It may be a very late series 1.1 before they switched to 2.x. I will check the serial next to ATX port when I go back home.
Anyway, I was just willing to help and could have done a VRM dump if it helped.
Ok just got home so I have a bit more info, that tends to confirm that this board was produced in a kind of transition period.
SN on the ATX connector is 2014xxx, so, even though it is a Rev 1.1, it was produced after the release of version HW 2.x. They did put an Avocent sticker even though it was produced the week just after Avocent support was fully dropped (end of March 2020), so probably one of the last series that got this sticker.
VRM are IR35201C747P and IR35204C740P (if my eyes are good enough) - (edit: I mistakenly reported the MOSFET reference initially (IR3555M) rather than the controller)
It has a 16MB only BIOS and the default BIOS was Naples only. It was used with a 7571 by the company who originally ordered it. I updated the BIOS with the Rome rbu, and tested it successfully with 7252 (8C) and 7452 (32C).
SN on the ATX connector is 2014xxx, so, even though it is a Rev 1.1, it was produced after the release of version HW 2.x. They did put an Avocent sticker even though it was produced the week just after Avocent support was fully dropped (end of March 2020), so probably one of the last series that got this sticker.
VRM are IR35201C747P and IR35204C740P (if my eyes are good enough) - (edit: I mistakenly reported the MOSFET reference initially (IR3555M) rather than the controller)
It has a 16MB only BIOS and the default BIOS was Naples only. It was used with a 7571 by the company who originally ordered it. I updated the BIOS with the Rome rbu, and tested it successfully with 7252 (8C) and 7452 (32C).
Last edited:
i need to know on which i2c slave addresses the VRM controller work if a Rome CPU is in socket.
in my not running Rome MZ31 shows the IR35201 at 0x24 but it is configured to 0x20,
and i2d addr. offset is set to 1 = NVM + offset from ext. resistor
so a resistor can change the i2c slave addr. and tell the VRM the CPU gen in socket (i guess)
but this is probablynot wired on my board. if i know the addr. i can adjust the NVM value to match the required addr. for Rome to work.
edit: not required. the addresses are already correct for rome, search for other issue.
Last edited:
Good luck with that. I´ve tried and got only boilerplate answers
@RolloZ170 with which CPU do you test Rome on your MB? Do you have a 7252 or 7452 too? I can take I2C traces for 201 and 204 controllers of the full boot, that would allow to compare any potential initialization issue, it would probably be easier to do the comparison with the same CPU to minimize the number of unknown.
What I can also share with you is a BIOS chip dump, as, even if the BIOS is the same, there will be differences in other areas (NVRAM). This may affect some of the board behavior (you will need to replace the MAC address of the network cards in my dump before flashing it).
Hetzner boards are the only one working based on this thread. And not even all of them.
I was looking at this offer: https://www.ebay.de/itm/188216075463?_skw=MZ32-AR0