PCIe bug with M2_1 slot on X870E Tomahawk

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
A similiar problem on MSI x870e EDGE Ti WIFI. M.2 slot 2 only operative on "auto". When change UEFI option to full speed m.2 without usb 40gb, ssd not recognized.
 
Upvote 0 Downvote
jlsmith.01712b902af said:
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!
Click to expand...
I have the SAME issue!!!

990 pro with speeds around 945 MB/s is so slow compared to 7000 MB/s WTF!!!!! The 9800 x3D cpu requires latest BIOS 2.A31 I think I still have time to take it back.

Switching from "auto" to "Gen 4" in BIOS does nothing. Samsung Magician software still says PCIe Gen. 1x4 no matter what I do in 2.A31 BIOS. I tried M2_4 slot and I'm getting PCIe Gen. 4x4 and speeds of 6387 MB/s read and 6224 MB/s Write
 

Attachments

  • SSDspeed.png
    SSDspeed.png
    62.5 KB · Views: 169
  • SSDgen.png
    SSDgen.png
    34.8 KB · Views: 108
  • SSDBIOS.png
    SSDBIOS.png
    18.3 KB · Views: 128
Upvote 0 Downvote
cgoodmanso159f02e1 said:
I have the SAME issue!!!

990 pro with speeds around 945 MB/s is so slow compared to 7000 MB/s WTF!!!!! The 9800 x3D cpu requires latest BIOS 2.A31 I think I still have time to take it back.

Switching from "auto" to "Gen 4" in BIOS does nothing. Samsung Magician software still says PCIe Gen. 1x4 no matter what I do in 2.A31 BIOS. I tried M2_4 slot and I'm getting PCIe Gen. 4x4 and speeds of 6387 MB/s read and 6224 MB/s Write
Click to expand...
Seems only the M2_1 and 2 slots have any issues. I found another thread on Reddit with people having the same problems.

https://www.reddit.com/r/MSI_Gaming/comments/1iexixz

I'm out of the return window, so I'll wait for a BIOS update or two to see if this gets resolved. If not, it's getting resold as is for a discount and I will tell the buyer why.
 
Upvote 0 Downvote
Joining the "I have the same issue!" gang here: 9800 x3d with latest BIOS revision on a MSI MAG X870E TOMAHAWK WIFI and I have the same issue of PCIe Gen 1 x 4 used for the 990 PRO Samsung Nvme drive. I really hope in a BIOS update that address this at this point.
Also tested changing "Auto" to "Gen4" but nothing, still PCIe Gen 1 x 4 reported and ~950 MB/s speed in benchmark.
 
Upvote 0 Downvote
Same issue! On two separate x870e Godlike motherboards on the latest bios, using a 9800x3d, and a Crucial T705. I just bought this motherboard and had some issues with it, this being one of them. I assumed it was faulty so I bought a second and planned to send the first back as faulty. But the second one has the same problem. And now I am reading all this, so, replacing the motherboard will not solve the problem... The M.2_2 slot works, but I have two M.2 drives I want running at full PCIe 5x4 speed... that is why I bought the Godlike to begin with; so simply not using the M.2_1 slot is not an option. Really hope this is a BIOS issue that can be patched...
 
Upvote 0 Downvote
It looks like Crucial identified this problem and it goes back to x670E boards too! AMD and MSI seem to be publicly pretending it is not a real issue. But I am very curious that the article says "MSI is the first manufacturer to respond with UEFI firmware updates that correct the issue" but... they didnt??? unless they only did on x670E motherboards? ... can we use an x670E bios on a x870e board? lol. details here: https://www.techpowerup.com/327243/...-gen-1-speeds-cause-boot-issues?cp=2#comments
 
Upvote 0 Downvote
Same problem. Brand new purchase from 2/17.
B2501 MAG X870E TOMAHAWK WIFI (MS-7E59) BIOS 2.A30 (was an issue before I updated as well... maybe A20?)
Samsung 990 SSD in M2_1 slot
Ryzen 9900X
 
Upvote 0 Downvote
Same problem. New build
MSI MEG X870E Godlike
AMD 9800X3D
7E48v1A24 Bios
M.2_1 Corsair MP700 Pro 2TB (drops randomly to Gen 4 speeds, or sometimes does not show in BIOS Has to be forced to Gen 5 in BIOS)
M.2_2 Sabrent SB-RKT5-4TB Seems to work ok when forced to Gen 5 (Auto setting gives random results)
 
Upvote 0 Downvote
Noticed the same issue on my new build. My drive is installed in M2_1 and is reverting back to 1x4, even when explicitly setting the gen in BIOS.
I've also had the drive fail to detect on cold boot, and had to restart from within Windows for it to show. The drive itself has no issues found by either MSI's BIOS self test or Samsung's test.

The image attached is from Samsung Magician and shows the detected Drive Max as 4x4, Slot Max as 5x4, but that the Interface is limited to 1x4.

This needs to be addressed.

MSI MAG Tomahawk x870e Wifi
7E51v1A24 Bios
AMD 9800x3d
Windows 11
Samsung 980 Pro 2TB m.2
 

Attachments

  • msi_m2_pcie.png
    msi_m2_pcie.png
    8.7 KB · Views: 164
Upvote 0 Downvote
A little more info for anyone looking at this thread. Even going into Windows sleep mode will have an effect on the performance and change data lanes it would seem.
Here are a couple reports from CPUz before and after coming out of a sleep mode seesion.

Controller NVM Express (NVMe) Controller at bus 2, device 0, function 0
Link Speed PCI-E 4x @ 2.5 GT/s
Volume c:\, 1862.3 GBytes (91.3 percent available)


Controller NVM Express (NVMe) Controller at bus 2, device 0, function 0
Link Speed PCI-E 4x @ 32.0 GT/s
Volume c:\, 1862.3 GBytes (91.3 percent available)

Running a CrystalDisk benchmark confirms the huge difference in performance that seems to be getting randomly juggled about.

I submitted a ticket to MSI with logs. I would suggest everyone else do it too.
 
Upvote 0 Downvote
Updated to BIOS 7E59v2A31 which was on the website, but not an update through MSI Center. Thus far enumerate at Gen 4x4 through 1/2 dozen sleeps/reboots/shutdowns. Seems better, but will continue to monitor.

UPDATE 2/25/25: BIOS 2.A31 does not resolve the issue. Reverted to Gen 1 after reboot today
 
Last edited:
Upvote 0 Downvote
kessler151402cf said:
Updated to BIOS 7E59v2A31 which was on the website, but not an update through MSI Center. Thus far enumerate at Gen 4x4 through 1/2 dozen sleeps/reboots/shutdowns. Seems better, but will continue to monitor.
Click to expand...
I'm on that same BIOS (same release date just another name) for the MEG X870E Godlike and have the problem.
 
Upvote 0 Downvote
I second everyone opening a support ticket if you have not already. This is very much a BIOS bug on M2_1. I opened a ticket and reported my observations so far.

1. With the NVME SSD in Slot 1, setting the M2_1 Link Speed to Gen5/Gen4 does not impact the link speed set upon boot into the OS. It almost always downgrades to 2.5GT/s, Width x2 (Gen1). Rarely, it may set to 16GT/s, Width x2 (Gen4) - I was only able to achieve this once.

2. With the NVME SSD in Slot 2, setting the M2_2 Link Speed to Gen5 for my M560 boots with a downgrade to 32GT/s, Width x2. Going back into BIOS and setting the USB4/M2_2 Switch to 'M2_2 X4' boots to 32GT/s, Width x4 (full Gen5 speed). This is set consistently with every reboot.

MSI MAG x870e Tomahawk Wifi (BIOS 7E59v2A31)
AMD Ryzen 7 9800X3D
G.Skill TridentZ CL30 (64GB)
MSI Spatium M560 (2TB, Gen5)
Bazzite (Fedora Atomic Linux distro)
 
Upvote 0 Downvote
j.william155102de said:
I second everyone opening a support ticket if you have not already. This is very much a BIOS bug on M2_1. I opened a ticket and reported my observations so far.

1. With the NVME SSD in Slot 1, setting the M2_1 Link Speed to Gen5/Gen4 does not impact the link speed set upon boot into the OS. It almost always downgrades to 2.5GT/s, Width x2 (Gen1). Rarely, it may set to 16GT/s, Width x2 (Gen4) - I was only able to achieve this once.

2. With the NVME SSD in Slot 2, setting the M2_2 Link Speed to Gen5 for my M560 boots with a downgrade to 32GT/s, Width x2. Going back into BIOS and setting the USB4/M2_2 Switch to 'M2_2 X4' boots to 32GT/s, Width x4 (full Gen5 speed). This is set consistently with every reboot.

MSI MAG x870e Tomahawk Wifi (BIOS 7E59v2A31)
AMD Ryzen 7 9800X3D
G.Skill TridentZ CL30 (64GB)
MSI Spatium M560 (2TB, Gen5)
Bazzite (Fedora Atomic Linux distro)
Click to expand...
I submitted a ticket tonight reporting the same issues. I’m guessing the response is a generic script one asking if you did this and that and then blaming user error. This needs to be taken straight up the chain of support for the dev team responsible of coding the BIOS update to be aware of it. Hopefully, with enough noise they should look into this.
 
Upvote 0 Downvote
Krisk7157302df said:
I submitted a ticket tonight reporting the same issues. I’m guessing the response is a generic script one asking if you did this and that and then blaming user error. This needs to be taken straight up the chain of support for the dev team responsible of coding the BIOS update to be aware of it. Hopefully, with enough noise they should look into this.
Click to expand...
I also submitted one a few days ago, and had a similar response. I have just been humoring their questions to see if I can bump it up to a proper issue, but so far I've shown that the issue is just the m2_1 slot, is not a problem with the drive, and that it occasionally boots with the proper speeds.

Been a couple days since I've had a response, though. I don't intend to close it until it's pushed up the chain or is confirmed as a bug.
 
Upvote 0 Downvote
Back
Top