PCIe bug with M2_1 slot on X870E Tomahawk

Status
Not open for further replies.
J

jlsmith.01712b902af

New member
Joined
Feb 16, 2025
Messages
14
So, the M.2_1 slot keeps reverting to PCIe 1 x 4 on a cold boot, sometimes just a reboot from within Windows. Manually setting it to Gen 4 or 5 within the BIOS will get it to run full speed until the next shut down/cold boot, then it's back to 1 x 4 again. Basically, I have to keep switching between gen 4 and 5 setting on every boot or it's stuck at 1 x 4. Auto is also broken. Samsung 990 Evo Plus 1TB.

I've seen another user on Reddit with the exact same issue. Same board, different drive. Anyone using this board, check your drive speed!
 
Sort by date Sort by votes
1160b02f1 said:
can that moderator help us out here and tell them they are full of it and tell them to just apply the same fix to the Godlike too?
Click to expand...
We do it almost weekly, The God board is a lot more complex than the Tomhawk, so it may take longer to find the problem.
 
Upvote 0 Downvote
MAG X870E TOMAHAWK WIFI + Samsung 9100 Pro

I've updated my BIOS with the last version (7E59v2A52) and it's seems that it fixed the problem for me. I've tested reboot and cold boot, my 9100 Pro remains in PCIe Gen 5 x 4 and bench shows good results :

Capture d'écran 2025-05-02 150231.png

Capture d'écran 2025-05-02 150222.png
 
Last edited:
Upvote 0 Downvote
rvillatoba14ec02d0 said:
The level of bullshit it takes to dare to respond that they "cannot replicate the issue"... While ALL buyers complain about it, and it's THE MOST EXPENSIVE MOTHERBOARD ON THE MARKET at around $1300. This may be the crux of the problem, it's probably so expensive that the number of buyers of the Godlike must be much lower than that of the Tomahawk. Basically, Godlike buyers are just a drop in the ocean of their customers, so they prioritize the cards they sell the most. I don't see any other reason for this nonsense. They just think the problem (complaining buyers) will go away on its own, so why bother?
Click to expand...
I would agree with this statement. I'd also say this is going to be the standard response now as they full well know there is a problem but will not acknowledge it. I guess they somehow believe this will protect them legally by not doing so. I've replied back to my ticket and waiting to get an answer back. I wish to try an exhaust all options before taking legal action.
 
Upvote 0 Downvote
Well, this is a bummer. Woke my CPU from sleep this morning, and CrystalDiskInfo and Samsung Magician are both reporting PCIe 3 x 4. I've actually never seen this speed reported; it was usually PCIe 1 x 4. Using the latest BIOS for the Tomahawk.
 
Upvote 0 Downvote
I know I've woken my CPU from sleep dozens of times since flashing the beta bios, and this is the first time it hasn't reported PCIe 4 x 4. Believe me, I check and benchmark regularly. :)
Maybe MSI can sprinkle just a little bit more of the magical dust they put into the beta bios and we can call it a day...
 
Upvote 0 Downvote
jonamar157e02dc said:
I know I've woken my CPU from sleep dozens of times since flashing the beta bios, and this is the first time it hasn't reported PCIe 4 x 4. Believe me, I check and benchmark regularly. :)
Maybe MSI can sprinkle just a little bit more of the magical dust they put into the beta bios and we can call it a day...
Click to expand...
Are you using the latest AMD chipset driver (7.04.09.545) or the one from the MSI website (7.01.08.129)? So far I have PCIe 4 x 4 constantly, whether after waking from sleep or rebooting, but I only installed the A52 version 11 days ago. Before that I used the A41 beta version for a month without any problems. Someone else has also reported a sudden reappearance of the bug after waking up from sleep (a few pages back), but that was with the beta version. No idea what triggers the bug for some people and not for others. The fix is probably not 100% perfect, the bug was still present for some people in Raid with the beta bios. I'm keeping my fingers crossed. So far, zero problems.
 
Last edited:
Upvote 0 Downvote
Alan J T said:
Yes, enable EXPO. It will set the correct Voltages for your RAM. All other options can be left on AUTO and do a reasonable job of getting good performance out of the CPU.
Click to expand...
isn't it concerning that EXPO enables a voltage that is highlighted as PINK/RED in the BIOS (indicating it is too high) ?
Why does MSI consider 1.4v high? (corsair 32gb 6000cl30 kit)
 
Upvote 0 Downvote
rvillatoba14ec02d0 said:
Are you using the latest AMD chipset driver (7.04.09.545) or the one from the MSI website (7.01.08.129)? So far I have PCIe 4 x 4 constantly, whether after waking from sleep or rebooting, but I only installed the A52 version 11 days ago. Before that I used the A41 beta version for a month without any problems. Someone else has also reported a sudden reappearance of the bug after waking up from sleep (a few pages back), but that was with the beta version. No idea what triggers the bug for some people and not for others. The fix is probably not 100% perfect, the bug was still present for some people in Raid with the beta bios. I'm keeping my fingers crossed. So far, zero problems.
Click to expand...
That was me a few pages back. Since that time, I never saw the bug again - and I tested multiple times a day in various scenarios (cold boots, sleep/wake, etc. etc.). Still have no clue how it was triggered other than just some random fluke when waking from that one particular sleep state. I actually went ahead and moved forward to the 2A52 BIOS update over this past weekend and I am running with the latest AMD chipset 7.04.09.545 drivers published on April 22. Fingers are crossed. So far things are smooth, but that doesn't mean anything, honestly. That recurrence with the 2A41 beta took nearly two weeks to show and up. Who knows... maybe this thing isn't 100% curable, after all.
 
Upvote 0 Downvote
rvillatoba14ec02d0 said:
Are you using the latest AMD chipset driver (7.04.09.545) or the one from the MSI website (7.01.08.129)? So far I have PCIe 4 x 4 constantly, whether after waking from sleep or rebooting, but I only installed the A52 version 11 days ago. Before that I used the A41 beta version for a month without any problems. Someone else has also reported a sudden reappearance of the bug after waking up from sleep (a few pages back), but that was with the beta version. No idea what triggers the bug for some people and not for others. The fix is probably not 100% perfect, the bug was still present for some people in Raid with the beta bios. I'm keeping my fingers crossed. So far, zero problems.
Click to expand...
I am using the latest AMD chipset driver. I wonder if that's why instead of dropping to PCIe 1 x 4, it dropped to PCIe 3 x 4? I'd never seen that speed reported before - it was only PCIe 1 x 4.
 
Upvote 0 Downvote
jonamar157e02dc said:
I am using the latest AMD chipset driver. I wonder if that's why instead of dropping to PCIe 1 x 4, it dropped to PCIe 3 x 4? I'd never seen that speed reported before - it was only PCIe 1 x 4.
Click to expand...
So we all use the latest bios with the latest AMD chipsets. Did the PCIe speed return to 4 x 4 after a reboot? Have you tried a cold boot or sleep again?
 
Upvote 0 Downvote
Yes and yes. After a restart it returned to 4 x 4, and this morning when I woke my computer from sleep it was 4 x 4.

To be clear, this is the first time in the literally dozens of times since installing the beta bios and now A52 that I've woken my CPU from sleep and it has reported something other than 4 x 4. The issue has gotten much, much better, but it seems there's still like a 1% chance it reverts to something else.

I actually take it as a positive that it reverted to 3 x 4 instead of 1 x 4. Since the only significant change to my system is the latest AMD chipset drivers, I wonder if that's partially related.
 
Upvote 0 Downvote
m153502d5 said:
isn't it concerning that EXPO enables a voltage that is highlighted as PINK/RED in the BIOS (indicating it is too high) ?
Why does MSI consider 1.4v high? (corsair 32gb 6000cl30 kit)
Click to expand...
64GB 6000/30 kit here as well, 1.4V isn't too much for this type of RAM
I think I'll skip the current BIOS (A34 for Godlike) due to the risk factor!
 
Upvote 0 Downvote
Personally, I never saw the PCIx bug on the Beta A41 for over a month of heavy usage and 7 days straight of a [***CENSORED***] load of testing and neither has it occurred on the A52 latest release using the recent chipset drivers. It could be, the PCIx bug is still there on the Tomahawk like a very small window of 2% to 3% of it occurring under a certain random trigger but I don't think it's worth time and effort now to get more reproduction rates, as you could be there for many hours a day and nothing will show up.
 
Upvote 0 Downvote
Alan J T said:
.
Click to expand...

Hi,
There's yet another bios out (vA6) for just the X870e Tomahawk
www.msi.com

MAG X870E TOMAHAWK WIFI | Gaming Motherboards｜Best Motherboard for AI PC｜MSI

The MAG X870E TOMAHAWK WIFI AMD AM5 gaming motherboard delivers cutting-edge connectivity, including Wi-Fi 7, 5G LAN, high-speed USB, DDR5, PCIe 5.0, and M.2 Gen5. Designed for AI-PCs and EZ DIY, it’s the top X870E ATX motherboard for AMD Ryzen 9000 CPU.
www.msi.com www.msi.com
Don't see one for other MSI X870/E boards.

Could you please help request proper release notes (like all other manufacturers provide)?
As I'm sure you'll agree, copy pasting notes from the previous bios doesn't help, and isn't useful. Kindly pass on this feedback, and appreciate your help.
 
Upvote 0 Downvote
Hi there.

I'm a X870 Godlike user, and was in another topic, following about the Dynamic Dashboard 3 issues, that looks like for now I'm not having.

But i'm having too the M2_1 speed problem, it in my tests, it run in PCIe 1x4 around 9 times on 10, and just rarely it work at 4x4 as may be. (Using a Samsung 990 PRO)
It work like this if the speed is on Auto in BIOS, or even forcing it at 4x4.

I'm so in contact with MSI with a long ticket that is open from like 30 days.
Initially they told me to update the BIOS with one coming from thrm, from a Google Drive link, and was not officially released on the X870E Godlike support page.
This BIOS was the 1A34, is still the last one today, I not installed it, before that appeared on the official page, so now I'm using it.
The BIOS seem to improve a bit the M2_1 speed, but not totally fix the problem.
I also noticed that this BIOS, fix the M2_2 that was listed as a removable drive, and now is not.

Sorry for the long introduction... Anyway...
Actually the MSI support, on the last ticket answer, is saying me that this issue is not fixable with a BIOS, and they are literally saying that is a problem that require a physical fix... To say it shortly, an hardware issue.
In total, in some days and answers they changed the things multiple times, saying opposite and confused things.

Is possible please to make here a bit more clarity on this problem? Because I'm actually not understanding how I may act, and if I may replace the board.
Actually I'm using the system drive in the M2_2 slot, that looks working always at full speed.

In my vision, if this M2_1 slot, sometimes work at 7300MB/s (4x4) but often it work at 900MB/s (1x4) it mean for me, the hardware is working, but the BIOS is not telling correctly how it may work on startup... Is it right or wrong?
I may think to replace the board? Or just wait for a BIOS that will fix it?

EXTRA QUESTION
I'm facing also an headache problem with Corsair RAMs RGB. That are working, but all the effects I set in ICUE, show every 3 seconds, a "block" in the animation, making it not smooth, and I can notice it even on the monitor into the preview.
All other devices managed by ICUE are just smooth and not have the issue.
I tried with a fresh Windows, with no drivers or anything, with just only ICUE installed, and I get the same thing.
But trying using SignalRGB, with this software the effects, or color changes, looks smooth, without any issues.
Looks like an ICUE actual release issue?

Thank you in advice for any answers on the M2_1 problem and this extra RAM RGB issue.
 
Upvote 1 Downvote
Status
Not open for further replies.
Back
Top