i have MZ31-AR0-HZ SN1829xx from itsp-24
Rome 7402P not working.
this is MZ31-AR0-00 1.1
i bought one from ram-koenig. SN1806xx - Rome 7402P not working.
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.
Rome 7402P not working.
i bought one from ram-koenig. SN1806xx - Rome 7402P not working.
That is why I shared the results on my board, which is not an Hetzner (it is a MZ31-AR0-00 V1.1) but still is Rome compatible. However it was produced very late (March 2020), so the date of production is really the determining factor. I will dump VRM registers over the coming weeks and generate I2C traces so we can debug the issue and try to backport the changes to older cards.
ok maybe other have more luck. some findings.
Serial and BMC MAC is stored in a serial EEPROM 24C256 or equivalent. (ATML H802 2ECL)
you can read this 24C256 with ch341a connected to the upper 24xx part of the socket.
inside is a sku.xml binary. save as sku.xml.gz and extract with gzip/7z/winrar.
00 02 00 = start offset of sku.xml.gz (0x0200)
29 0A 00= lenght of.... (0x0A29)
if you want to connect to VRMs:
there are three(3) i2c TMP75 thermal sensors on the board. they are one the same bus. i connected with a SOP8 clamp.
Location of the two chips and pin1
IR35201 is Vcore
IR35204 is VIO(SoC)
and two IR3584 for RAM
Serial and BMC MAC is stored in a serial EEPROM 24C256 or equivalent. (ATML H802 2ECL)
you can read this 24C256 with ch341a connected to the upper 24xx part of the socket.
inside is a sku.xml binary. save as sku.xml.gz and extract with gzip/7z/winrar.
00 02 00 = start offset of sku.xml.gz (0x0200)
29 0A 00= lenght of.... (0x0A29)
if you want to connect to VRMs:
there are three(3) i2c TMP75 thermal sensors on the board. they are one the same bus. i connected with a SOP8 clamp.
Location of the two chips and pin1
IR35201 is Vcore
IR35204 is VIO(SoC)
and two IR3584 for RAM
Last edited:
you login to BMCWebGUI
Maintenance, Firmware Update
select Firmware image (image.RBU)
click start Firmware update
change update type to "BIOS"
Proceed to flash
the 8144 update is included in any BIOS Rome/Milan from R03 and up (AMI ROM Hole)
and written volatile at startup in the VRMs (all four)
the VR update 8144 is not the thing to make a non Rome supporting motherboard work with Rome.
the update is only cosmetics.
I have updated my MZ31 with the register set of the 8144 update with USB005(DIY) + PowIRCenter and stored to NVM,
I read back and compared with update result in 100% match, but no change, no Rome(tried many BIOS R03,R18,R34,R40).
( pls note the IR35201/IR35204 have only 5 OTP banks left to program )
edit: i got no PMBus faults reported.
edit: Naples 7401P still work after VRM update (w. Fxx BIOS of course)
I have done that several times already and with different BIOS-es (MZ31_flash_R34.RBU, R40_F23, R18_F18), even cleared BIOS with jumper . The guy I asked finally refused to test my 7502P in his mobo, so I have borrowed another 7252 CPU. The result is completely the same. No POST, just CPU presence detected message in BMC. No power controls working on my MB SN 1828xxx for 7502 nor 7252 and any BIOS I mentioned before. I think I am done with trying, if anything else will not appear from your and @nfa findings.
The only difference I have found when AMD 7401P inserted and system POSTed and worked as expected, two more LEDs near BMC LED switched on. The label says LED_PE1 and LED_DL1. Haven't found any description for those in the manual.
I was confused if any other BIOS than original F11b is working in my MB so I tried F23. System POSTed with my original 7401P an F23 as well.
I do not understand why my MB even not responding to power controls unless 7401P and Naples BIOS flashed.
[Update]
It seems my power switch was malfunction-al, at least right now with 7401P inserted, shorting pins 11&13 working as expected.
Maybe I will try resetting when turned on with shorting pins like recommended by @RolloZ170 and ROME cpu during the weekend once more to get some POST codes using bioscode utility.
Last edited:
no way. he shows so many boards and sells one one.
闲鱼 - 闲不住？上闲鱼！
闲鱼，中国领先的闲置二手交易平台，趣味生活社区，从最初的C2C交易，到用户在这里分享技能、兴趣与经验，从物品到服务、从交易到交流，如今闲鱼累计用户数超5亿，日均交易额已突破10亿，每天都有400万件闲置物品在平台上发布。同时，闲鱼以闲置交易为起点，还发展出一键转卖、极速回收、验货宝、闲鱼小法庭等服务；并在2024年闲鱼成立十周年之际推出了闲鱼简历功能，通过闲置时间、技能来赚钱，让你的人生另起一行。
www.goofish.com
Last edited:
Thought so.
Just don't get it what it is his angle here. If you don't get the goods you can always dispute the deal and get money back. So it's just stupid games.
in standby (just PSU on, CPU off) i see only the RAM IR3584 VRMs.
the IR35201 & IR35204 comes up only after power on.
can you check if the IR35204gets 3.3V on your board in standby ?
measure point is here on SMD capacitor to any ground e.g. LAN port shield or USB header pins.
update:
on my MZ31-AR0 1.1, the VCC of IR35201(Vcore) and IR35204(VSOC) comes from 24pin ATX 3.3V source.
Last edited:
@RolloZ170 for now cancelled my mz31 board, but I am still watching this thread. With your determination I am quite sure you will crack this problem!
Unfortunately no luck at all. I have switched my front panel buttons for working ones. The behavior with ROME CPU and any ROME BIOS is the same as ROME CPU and NAPLES BIOS, MB doesn't respond to power nor reset button and BMC reports "Processor presence detected" (even for ROME CPU and Naples BIOS combination) as well. I have tried even more bios files (flash_R.rom converted to RBU with checksum extension) like server_bios_MZ31-AR0_R03_F16.zip, server_bios_MZ31-AR0_R09_F18.zip, server_bios_MZ31-AR0_R12_F18.zip and server_bios_MZ31-AR0_R12_F18.zip ....
I think I am done right not with more trying unless @RolloZ170 or @nfa find the solution to move a little bit forward.
@RolloZ170 do you think there is a difference when flashing BIOS from BMC and from UEFI? The Gemini thinks so
but I don't. I couldn't find ".img" to try it anyway.Just FYI, even with ROME BIOS and Naples CPU, power button works as expected. When I put ROME CPU, neither BIOS (ROME or NAPLES) responds to power button. Strange, but it is how it is :-|. I hope I didn't have bad luck for locked ROME CPUs, both of them (7252/7502).