I wonder if perhaps there is a way to force a more complete flash of the BIOS file. I think the ME firmware is only updated if it is older than that in the BIOS file.

There are ways to manually update the ME firmware. I think it involves editing an existing BIOS file and injecting the newer firmware into it. Then flashing that modified file. Of course this is complex and risky.

swaaye wrote on 2024-11-21, 18:56:

I wonder if perhaps there is a way to force a more complete flash of the BIOS file. I think the ME firmware is only updated if it is older than that in the BIOS file.

There are ways to manually update the ME firmware. I think it involves editing an existing BIOS file and injecting the newer firmware into it. Then flashing that modified file. Of course this is complex and risky.

Well from the Googling i've been trying to do - I can't tell if the ME Firmware is located on the CPU, or it is a chip on the motherboard.
And I made the mistake of not remembering what BIOS the board had before I flashed it to the latest. But in that regard, flashing the BIOS didn't fix the ME issue unfortunately.

The ME engine is a bit of a mystery for sure.

Maybe you could try booting from the backup BIOS and see if the board behaves differently. I've never done it myself but I found these tips:
https://www.reddit.com/r/gigabyte/comments/6b … _bios_i_have_a/

It may also be possible to flash an older BIOS.

Have you tried putting a new battery in it?
Maybe it's just me, but newer motherboards (~2010 onwards) have a habit of acting very oddly when the CR2032's dying instead of just plainly stating that it's flat.

Might be worn out flash chip.

First: The Intel ME version is not directly tied to a specific BIOS version, sometimes the ME firmware is updated after a few BIOS updates, sometimes every BIOS update depending if Intel released a new ME firmware about same time as the new bios and the OEM wanted to include that newer firmware. It is not in the main bios eeprom chip but in a eeprom in the chipset (or in a secondary small eeprom directly connected with some chipsets iirc)
Brief on what it is: https://www.intel.com/content/www/us/en/suppo … t-software.html
Second: I never install the Intel Management System software to the OS ever since having OS issues with a DQ77MK after installing it years ago after a BIOS update. Windows 7 and up has built in drivers for the Intel ME access but does not really do anything with it unless you install the software. It really serves no purpose under general home use/game use much like the Intel TPM.
As for why the ME keeps rebuilding: Is it rebuilding because you installed the ME software to the OS ? Have yet to see a motherboard rebuild the ME on a bootup from BIOS, but guess it is possible, no on second thought not.

swaaye wrote on 2024-11-21, 19:20:

The ME engine is a bit of a mystery for sure. […]
Show full quote

The ME engine is a bit of a mystery for sure.

Maybe you could try booting from the backup BIOS and see if the board behaves differently. I've never done it myself but I found these tips:
https://www.reddit.com/r/gigabyte/comments/6b … _bios_i_have_a/

It may also be possible to flash an older BIOS.

Thanks for those tips. I might try them. I have definitely already tried flashing an older BIOS, went back to the first release. It did not help remedy the problem.

jmarsh wrote on 2024-11-22, 00:37:

Have you tried putting a new battery in it?
Maybe it's just me, but newer motherboards (~2010 onwards) have a habit of acting very oddly when the CR2032's dying instead of just plainly stating that it's flat.

I could try that. I did take out the battery completely to see if that would help at all. Unfortunately it did not.

Horun wrote on 2024-11-22, 03:42:

First: The Intel ME version is not directly tied to a specific BIOS version, sometimes the ME firmware is updated after a few BI […]
Show full quote

First: The Intel ME version is not directly tied to a specific BIOS version, sometimes the ME firmware is updated after a few BIOS updates, sometimes every BIOS update depending if Intel released a new ME firmware about same time as the new bios and the OEM wanted to include that newer firmware. It is not in the main bios eeprom chip but in a eeprom in the chipset (or in a secondary small eeprom directly connected with some chipsets iirc)
Brief on what it is: https://www.intel.com/content/www/us/en/suppo … t-software.html
Second: I never install the Intel Management System software to the OS ever since having OS issues with a DQ77MK after installing it years ago after a BIOS update. Windows 7 and up has built in drivers for the Intel ME access but does not really do anything with it unless you install the software. It really serves no purpose under general home use/game use much like the Intel TPM.
As for why the ME keeps rebuilding: Is it rebuilding because you installed the ME software to the OS ? Have yet to see a motherboard rebuild the ME on a bootup from BIOS, but guess it is possible, no on second thought not.

I can only assume the previous owner of the board did have the ME drivers installed. I have zero history of the actual owner, since I got this from a middleman who never tried using it himself. He simply took the power supply and RAM from the computer this motherboard/CPU combo was from.

I will say some positive news about stability: I left the computer running Unreal Tournament 2004 in software rendering, spectating a match, for 10 hours while I was at work. The computer was still running perfectly fine when I got home. So definitely safe to assume the CPU checks out ok.

Horun wrote on 2024-11-22, 03:42:

It is not in the main bios eeprom chip but in a eeprom in the chipset (or in a secondary small eeprom directly connected with some chipsets iirc)

That doesnt sound right. I repaired a ton of never stuff (mobos, laptops) and its all in one main flash chip. ME reinitializing might be because of ME region data mismatch or worn out flash chip.
https://github.com/platomav/MEAnalyzer
https://github.com/LongSoft/UEFITool

ThruMy4Eyes wrote on 2024-11-22, 05:08:

I will say some positive news about stability: I left the computer running Unreal Tournament 2004 in software rendering, spectating a match, for 10 hours while I was at work. The computer was still running perfectly fine when I got home. So definitely safe to assume the CPU checks out ok.

Another owner seemingly fixed this same problem with your board...

https://www.tweaktownforum.com/forum/tech-sup … a-h67ma-usb3-b3

using info from this thread...

https://hardforum.com/threads/tools-to-flash- … or-ftk.1726429/

if you have spi chip programmer and you read your bios i can insert clean ME region.
when i used to work on laptop repairs that was a common issue that ME region gets corrupted.

rasz_pl wrote on 2024-11-22, 07:03:

That doesnt sound right. I repaired a ton of never stuff (mobos, laptops) and its all in one main flash chip. ME reinitializing might be because of ME region data mismatch or worn out flash chip.
https://github.com/platomav/MEAnalyzer
https://github.com/LongSoft/UEFITool

You are right, was confusing the built-in TPM thing on my dual server board.
I think the eeprom is damaged per reading some Intel docs on ME and its firmware should be read only like main bios (if I read the docs properly).
This boards bios the ME firmware starts at 3000h, confirmed by comparing the FW.bin. Interesting thing is that its last BIOS upgrade for this had a set of Intel ME tools (Intel FWUPD and MEinfo) in subfolder including a 2Mb ME FW.bin file.
Still will not install the ME software to any after the issues with that Intel Q77 board....

added: just checked the F2 bios ( F2, 1.46 MB, Mar 24, 2011, Update ME size to 256K ) and compared to F8.
It has same ME firmware starting at 1000h. In fact the whole top from 0h to 200000h are identical. Main bios starts at 200000h. So flashing the F8 should have re-wrote the ME firmware and fixed it per this if the eeprom/fwh is fully functional:

PC Hoarder Patrol wrote on 2024-11-22, 11:44:

Another owner seemingly fixed this same problem with your board...
https://www.tweaktownforum.com/forum/tech-sup … a-h67ma-usb3-b3

Note: just read that after flashing with ME update or a re-write one should turn off the computer and PSU or pull power cord so no 5vsb is present in case it interferes with proper initialize/reset of the ME portion...
also said : if not updating ME should not have to pull power... hmmm

There is ME firmware and then there is ME data (dynamic, often written to) in the ME region. If this data is corrupted board will have problems with long boots and random shutdowns, in that case one needs to clean ME region and let ME firmware reinitialize it (takes loon boot or even couple reboots with blank screen before its done). But if flash is worn out it will keep corrupting itself.

Some time ago, I prepared some HP 8300 for my friends and they have IME.

On their support page, they list updates for IME , on sp80282.exe. That pack contains DOS based tools to flash ME firmware and the firmware itself, not only for that 8300 but for earlier and later incarnations of that. I don't know if that tools only works for HP BIOS or can flash other boards firmware. Maybe some of that tools can also erase your IME data area.

Also, on that HP 8300 you could disable IME. I don't know if it only "hides" the devices so they don't show in Windows, or stop the thing entirely. Have you checked if your board can disable IME? If so, does it change your boot issues?

rasz_pl wrote on 2024-11-23, 07:35:

There is ME firmware and then there is ME data (dynamic, often written to) in the ME region. If this data is corrupted board will have problems with long boots and random shutdowns, in that case one needs to clean ME region and let ME firmware reinitialize it (takes loon boot or even couple reboots with blank screen before its done). But if flash is worn out it will keep corrupting itself.

Yeah this definitely sounds like the issues this is having!

Zup wrote on 2024-11-23, 08:46:

Some time ago, I prepared some HP 8300 for my friends and they have IME.

On their support page, they list updates for IME , on sp80282.exe. That pack contains DOS based tools to flash ME firmware and the firmware itself, not only for that 8300 but for earlier and later incarnations of that. I don't know if that tools only works for HP BIOS or can flash other boards firmware. Maybe some of that tools can also erase your IME data area.

Also, on that HP 8300 you could disable IME. I don't know if it only "hides" the devices so they don't show in Windows, or stop the thing entirely. Have you checked if your board can disable IME? If so, does it change your boot issues?

I've been through the BIOS, and Gigabyte doesn't have any IME settings that I can think of. The next time I have it up on the table I will give it another look.
It's a shame their stupid ME causes these problems, because otherwise the CPU does all the CPU things it should in Windows perfectly fine!