- Joined
- Mar 6, 2013
- Messages
- 266
- Motherboard
- Gigabyte X299X Designare 10G
- CPU
- i9-10980XE
- Graphics
- AMD 6900XT
- Mobile Phone
Superb work, @JTR ! Working well for me on BIOS F3C. Thanks so much for all the time you spent diagnosing and fixing, and for sharing the fix with us in a public repo.
Came at a great time for me, because just the other day I realised that with my Intel X520 fibre 10GBe NIC installed, I was getting the issue on every boot. The NIC adds a multi-second delay to BIOS initialisation - related to its UEFI driver I suppose - and this was enough to cause every boot to trigger the safe mode reboot, including from SSD/NVMe which had always worked for me before.
Fingers' crossed that Acidanthera will accept it, once you're ready to submit. Looking at the code it looks straightforward enough that hopefully they'll be willing to.
Re GB, sadly this does not surprise me. I swear this must be the buggiest motherboard I have ever seen. I just can't understand how they can have the audacity to not patch any of these issues after 12+ months.
I finally have RAM overclocking working thanks to that video you linked. The workaround is easy, but until you know how it's easy to assume that your RAM just won't OC, or that all RAM OC is broken on this board. And it's a bug they've definitely known about for at least 13 months and apparently don't give a damn about fixing.
On top of that there's an issue with VCCSA resetting back to AUTO (or at least to 0.9V). Again only affects overclocking, but this is a £600+ motherboard that's marketed for "pursuing the highest overclock performance". I saw this issue a month or two ago and at the time thought it was macOS specific, but it's not. Whenever I set VCCSA ("CPU System Agent Voltage" as it's called in the BIOS) to a specific value, I will find later, after shutting down in Windows or macOS, that it's back to 0.9V. Then to make matters worse, it can't just be re-applied by hitting F10 in the BIOS. In order for it to recognise you made a change, you have to alter the value. Eg if it was set to 1.250V, you need to set it to 1.245 or 1.255, so that F10 recognises a change has been made and sets it. Meaning if you wanted it at a specific value, you'd have to do TWO reboots and F10s. Then next time you shut down it'll revert again and you have to do it all again. This bug could actually be dangerous, because I've read that when doing RAM OC, VCCSA should never be left at auto as it can apply too-high voltage levels.
Then there's a less serious but still stupid and annoying bug: if you set DRAM speed to any non-default value, the BIOS shows that setting as changed every time you enter the BIOS in future. So if you go into the BIOS and change nothing, it will still always prompt you to save your changes because it thinks DRAM speed has changed. This effectively breaks Boot Override completely, because it won't let you boot override if you've made BIOS changes - it prompts you to save them and reboot instead. Not nearly as big a deal as the other bugs, but still dumb and yet another thing that can waste the user's time.
All that's on top of occasional BIOS resets when trying to apply profiles, and the French language issue which I now seem to get regularly for some reason. I hadn't had it go into French for ages, and thought it only happened after boot failures or BIOS resets. But now it seems to be happening every time I shutdown in Windows, making me think it's related to using the F12 boot menu, as that's how I'm currently booting Windows, rather than doing it from OpenCore. What a mess.
Once I'm on a stable config I'm going to double check these issues in F3B so I can decide between the versions. I know the RAM training voltage issue is in both, but not sure about the other issues such as VCCSA.
Last edited:
