MAG X670E Tomahawk wifi starting... too fast?

Spagettifritz

New member
Joined
Feb 26, 2024
Messages
15
Hey guys,

As some of you probably know, MSI recently released the update 7E12v18 for the X670E Tomahawk Wifi, which updates AGESA and significantly shortens boot times with "Memory context restore" enabled.

So far so good, the boot times have improved drastically for me, before I had a post of 60 seconds, now less than 2 seconds! Yay!

However, I have now encountered another new problem.
I have 3 PCIe expansion cards, a 2080 Super, a Mellanox Connect-X3 10 GBit/s NIC and an Elgato 4k60 Pro Mk.2 capture card.
Before the BIOS update, everything worked perfectly. Now, with activated "memory context restore" in the latest BIOS version and the short boot times, the Elgato Capture Card is constantly missing in the system.

Only a warm reboot helps. But that's where it gets interesting:
- Cold boot -> card is not present in Windows
- Warm boot -> card is present
- Hibernate & start -> card is not present
- Sleep & start -> card is not present
- Cold start, but first into BIOS, then warm restart without changes -> card is present

What was tried:
- CMOS reset -> without success
- Capture Card tested in other PCIe slots -> without success, the error moves with the card, it is not due to the PCIe slot
- With every warm boot, the card is present and functional again
- Delay post with 3 Seconds "fake" post duration (this is an option on some boards...!) -> The BIOS of the X670E does not have that option - at least I couldn't find it.

My thoughts:
I think that the Elgato capture card has a firmware that doesn't start as fast as the PC starts from shutdown/hibernate/sleep, because it apparently does not get recognized.
Now that I'm happy not having a space rocket launching (GPU fans on 100% for 60 seconds) while booting my PC with the old, long boot times, I have a reproducible error with one of my PCIe cards.

Is anyone else having this issue with the new BIOS Update?
Is this a case for MSI to "artificially extend" boot times to give the PCIe cards a chance to initialize?
Or implement the "fake" post delay, where one can set a duration from 0 to 10 seconds delay in BIOS?
I'd appreciate every thought and suggestion of you guys.

Here is a summary of the data:
MB: X670E Tomahawk Wifi
CPU: Ryzen 7 7800x3D
RAM: 2 x Kingson DDR5-6000 32 GB RAM (HCL compatible) with A-XMP enabled for 6000 MHz
BIOS: 7E12v18
OS: Windows 10 x64 Pro

Thank You!
 
Got a response from Elgato Support.

The fact that the AverMedia card and driver installed didn't resolve the issue with capturing cards shows a much larger issue.
They recommend to let MSI know (as they are the MB manufacturer), that neither the Elgato nor the AverMedia card worked.

Each BIOS update is supposed to improve Memory Stability
I see, I see...

But well ya:wall::wall::wall::wall:
:giggle:

Do you have Memory Context Restore set to Enabled
Power Down and Geardown in Advanced Memory setting set to Enabled
And also set Ignore memory change to Enabled
Will try this.

Thanks.
 
I don't suppose you have a spare non-X3D CPU by chance.
Oh, unfooooortunately not...
But if MSI covers the costs, I'd be happy to test!! :hiha:

Jokes aside... You thinking about a malfunction with the (my?) x3D CPU? Or rather that the MB BIOS doesn't work correctly with some x3D configs?

I also created a support ticket with MSI official support now, as I want this to be fixed and solved. Ticket Number: #1203271
(Greetings to MSI support staff!)
 
With the sheer number of problems and over 90% of them seem to be X3D CPU, I have to wonder if they are up to doing as advertised.
I didn't know anything about that yet.
But I still think that if the CPU had a problem, that some other errors would also occur.
The gaming performance is not only OK, but actually great and there are no permanent problems with the RAM training so that the computer would not start at all, it only happens after restarts when the system has been running for a longer period of time.

As for the problem with the capture card... I have no idea.
Elgato support is currently offering me to replace the capture card with an external 4K X... that won't solve the problems with the PCIe card, but rather avoid troubleshooting and will create other problems for me if I have to go via USB.

This XKCD explains it best
:ROFLMAO:, but also true.
 
Hey all,

MSI Support replied to the Elgato capture card issue:

"This problem is related to Elgato 4k60 Pro Mk.2 capture card. We are working on the issue with AMD, it may take some time to fix the issue. Sorry for the inconvenience caused you!"
Hoping that this won't take another year to fix...

I also asked about the RAM reboot issues...
...after being asked to test with 1 RAM module, then with 2, then with A-XMP on/off

waiting for a reply.
 
OK guys, so the recent reply from MSI is as follows:

Regarding the Elgato capture card issue:
"For the card issue, sorry to say that we may not release new officially BIOS for it as it may cause other PCIE compatibility issue. We think that your current compatibility issue cannot be fixed by BIOS update currently."
... but with the new beta BIOS (7e12v1i3 from 2025-04-29), I also haven't had any more issues with the capture card... still couldn't test it over a long time, but it seems to be good now.

And for the reboot problem, the current solution seems to be the latest beta BIOS (7e12v1i3 from 2025-04-29) with...
- memory context restore option set to "ENABLED"
- NO profile (A-XMP / EXPO) activated

Haven't had a problem with reboots for over a week now.
Not sure, how the memory context restore option can do something when there is no profile to "restore" to. Asked support twice about this, no answer.

Also, current recommendation is to NOT activate an A-XMP or EXPO profile...
"You can just keep the current settings if the restart function is working normally now."
This makes no sense. I paid for A-XMP / EXPO, I wanna use A-XMP / EXPO.... I'll test it anyway...

Just realized, that there is also a new BIOS, the non-beta version of 7e12v1i3, released 2025-06-10...
Gonna test this as well, I guess....

Cheers.
 
Hey guys,

Sorry for being silent for over a year. I just wanted to give you an update on this case.

Over the past several months, I've tested a number of BIOS versions, but none of them have really satisfied me.

I still can't use my 10 Gbit/s network card because it continues to cause issues with both the NIC and the capture card. As a result, my system is currently running with only the graphics card and the capture card installed.

The capture card itself is currently working flawlessly. However, because of my previous experience with the recurring motherboard issues, I hardly use it anymore. At least the system itself is completely stable with the capture card installed.

Unfortunately, more and more RAM-related issues have appeared instead. As soon as I enable A-XMP or EXPO, I run into another problem about once a week: the system will no longer wake up from Sleep (S3). The motherboard's debug LED indicates that the CPU is detected correctly, but then it gets stuck during RAM training.

This is not the same issue as before, where RAM training simply took an unusually long time. In this case, the system does not resume from Sleep at all during that boot attempt. The only way to recover is to completely power the system off and then turn it back on.

However, I also found a solution...
Disable A-XMP / EXPO. Leave the RAM running at its stock configuration.
That's it. With A-XMP / EXPO disabled, the system boots and resumes without any issues, whether it was completely powered off, in hibernation, or in Sleep (S3).

So, to sum it up:

None of the BIOS updates have improved the overall situation.

I've now disabled A-XMP / EXPO permanently, and after everything I've experienced, I don't even want to use it anymore. It simply makes no sense if MSI - or AMD, I honestly don't know where the root cause lies - can't get this issue under control.

Users with ASRock motherboards have told me that they experience similarly long RAM training times, but they don't have any issues with waking their systems from Sleep or Hibernation.

The overall quality of this hardware generation has been truly disappointing.

I'm leaving this thread open so other users can contribute their findings or until some moderator decides to wrap it all up and close it.
Thank you to everyone trying to help, especially @Alan J T -

Maybe it's time for me to start looking for a motherboard manufacturer that can deliver reliable BIOS updates and doesn't struggle with RAM overclocking, expansion card compatibility, and basic boot stability years after launch.

Cheers.
 
However, I also found a solution...
Disable A-XMP / EXPO. Leave the RAM running at its stock configuration.
That's it. With A-XMP / EXPO disabled, the system boots and resumes without any issues, whether it was completely powered off, in hibernation, or in Sleep (S3).
I don't use Expo. I have set the memory timing and speeds manually in the BIOS I find it to be more reliable this way with my configuration of 4 Ram Sticks
 
Have you tried enabling EXPO but then manually lowering (override) the "CPU NB/SoC voltage" to 1.25 Volt?
The MSI BIOS is setting this voltage to 1.3 V when enabling XMP/EXPO or enabling the "Optimized Performance Profile". This higher voltage is known to cause issues for warm restarts and is fixed by manually setting it to 1.25 or lower.
This workaround might also fix the sleep issue.
 
Back
Top