PCIe bug with M2_1 slot on X870E Tomahawk

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!
 
Have the same PCIe 1.0 x4 issue on T705 in a MAG X870E Tomahawk with latest BIOS. I did have PCIe 5 briefly earlier today on the M2_1 slot, but that went away on reboot.
 
Just joining the party today. I've built my system few weeks ago and been fighting various problems, but finally got to this one. I'm using Samsung 9100 Pro with X870E Tomahawk and latest BIOS (7E59v2A4). Generally speaking I get the full speed in a reboot or two. Today I was talking to Samsung support to see if it may've been their problem, but just found this post. I've tried today moving my secondary nvme (990 Plus Evo 1 Tb) to the m.2_1 and while it does not drop to gen1, it fluctuates between gen4 x 4 and gen5 x 2 pretty much every boot. It does not change the performance of this particular driver, but for 9100 it goes between gen1, gen4 and gen5, like others.
 
I am also having this problem. I built my system on 3/3/25. Sometimes a single reboot will fix it and other times its just frustrating because it never switches back to full speed. One thing I've tried that seems to work *sometimes* is unplug any USB device and reboot. I've had more success with this method (It may be by chance or something else).
 
Last edited:
Until MSI issues a new official release with the fix (7E59v2A4 does NOT address the issue), anyone who visits this thread in search of a solution on the MSI X870E Tomahawk WIFI board has two options:

1. Install the beta BIOS (v2A41) from the Google Drive link mentioned in [this previous comment].
2. Move your drive to Slot 2 and turn off the USB sharing for that slot to ensure full speed for the drive.

Note: Some members have reported inconsistency with RAID speeds on the beta firmware, so YMMV if you have a RAID setup. I have had no issues with my single drive installation in Slot 1 on the v2A41 beta BIOS.
 
Just joining the party today. I've built my system few weeks ago and been fighting various problems, but finally got to this one. I'm using Samsung 9100 Pro with X870E Tomahawk and latest BIOS (7E59v2A4). Generally speaking I get the full speed in a reboot or two. Today I was talking to Samsung support to see if it may've been their problem, but just found this post. I've tried today moving my secondary nvme (990 Plus Evo 1 Tb) to the m.2_1 and while it does not drop to gen1, it fluctuates between gen4 x 4 and gen5 x 2 pretty much every boot. It does not change the performance of this particular driver, but for 9100 it goes between gen1, gen4 and gen5, like others.
You won't be getting Gen 5 x2 on a 990 Pro! It is interesting that it's not dropping to Gen 1 on the non-system drive though.
The 9100 will always run at Gen 5 on the second M2 slot.
Apologies if I mis-read your post
I have my 9100 in M2_2 now.
 
I need to buy a new mobo to mount my new configuration with a 5090 and I'm considering the x870e, I learned about this problem with the m2.2_1
I need to use 2 m2 and 4 sata, using m2.2_2 and m2.2_3 or maybe m2.2_3 and m2.2_4 would I have problems with halving the bandwidth on the GPU line or other side effects?

Do msi x870 chipsets also have this problem?

p.s.

actually I was oriented towards a b850 tomahawk, which does not seem to share the m2 lines but I know it has problems with the lan that disconnects, I don't know if they have solved it

thanks
 
Last edited:
You won't be getting Gen 5 x2 on a 990 Pro! It is interesting that it's not dropping to Gen 1 on the non-system drive though.
The 9100 will always run at Gen 5 on the second M2 slot.
Apologies if I mis-read your post
I have my 9100 in M2_2 now.
Scratch that, it did drop to gen 1 once out of 30 or so reboots, so it's much more stable, also it's 990 Evo Plus, not 990 Pro. 990 Pro is gen 4 x 4 only afaict, I got 990 Evo Plus specifically to run it in m.2_2 in gen5 x 2 mode so I can have other x 2 for USB4. 9100 Pro did run fine in gen.5 x 4, but the point is not to lose USB :) Also I use 990 Evo Plus for the system install and 9100 for "data"
 
I've also had the problem with speeds, and I downloaded the Beta BIOS A41, which solved the problem (just normal flash BIOS without CMOS reset).

I have two Samsung 990 PRO drives, one on the M2_1 port and the other on the M2_2 port. With USB 4.0 disabled, both M.2 drives now show as 4x4 in Samsung Magician, and the tested speeds are good.

No problems with the beta BIOS either. The only thing I had to do was manually set the RAM speed to 6000 MHz because after the update, it was reset to 4800 MHz.

Hope this update will work for everyone.
An update: Sometimes, when booting the PC, my second M.2 drive doesn't show up. Additionally, after a simple sleep or log-out, when I sign back in, the second drive (M2_2 - the one for games, movies, etc.) doesn't appear, and I have to restart the PC for it to show up.
 
Until MSI issues a new official release with the fix (7E59v2A4 does NOT address the issue), anyone who visits this thread in search of a solution on the MSI X870E Tomahawk WIFI board has two options:

1. Install the beta BIOS (v2A41) from the Google Drive link mentioned in [this previous comment].
2. Move your drive to Slot 2 and turn off the USB sharing for that slot to ensure full speed for the drive.

Note: Some members have reported inconsistency with RAID speeds on the beta firmware, so YMMV if you have a RAID setup. I have had no issues with my single drive installation in Slot 1 on the v2A41 beta BIOS.
Just want to chime-in on this because I can confirm that beta 2A41 does not 100% solve the issue, but it is definitely a major improvement and I highly recommend it. As I posted on March 27 after about 10 days using the beta, it is still possible for the M.2_1 slot to screw up in some interesting ways (woke my system up from sleep to find my Samsung 990 Pro reporting 1x4 connection but testing at 4x4 and 3x4 speeds). After numerous cold boots and restarts, I resorted to clearing CMOS and have not had it show up again on 2A41 since March 28.

Still no clue what could have triggered it (all this machine does is play Monster Hunter Wilds), but I continue to log all of my cold boot, restart, and wake activities to track if it happens again on the 2A41 BIOS.
 
Maybe it's time to start pushing this issue to the bigger known YouTube reviewers to show the issues of these boards. Maybe that publicity will light a fire under someone's seat to get some action here. So far on the Godlike board it has been crickets!
Does anyone have a good suggestion on who and how? i sent an email to GamersNexus a few weeks ago but got crickets. Considering how many emails he may get and that a lot of his audience is probably not completely tech literate yet (some people are watching youtubers to learn), i would not be surprised if he just assumed I am a crazy person (maybe I am a crazy person, but I am also an engineer with a M.S. and years of experience with root cause analysis and building PCs) so maybe he does not take all his emails seriously. BUT - A 100% defect rate on all MSI x670e and x870e motherboards that MSI had failed to address for multiple generations now is a pretty damn big deal. Honestly, it impacts a lot more people than a lot of the other stories some of these YouTubers decide to cover. Maybe we need multiple people reaching out so they dont just ignore it in the flood of emails they might get.
 
Does anyone have a good suggestion on who and how? i sent an email to GamersNexus a few weeks ago but got crickets. Considering how many emails he may get and that a lot of his audience is probably not completely tech literate yet (some people are watching youtubers to learn), i would not be surprised if he just assumed I am a crazy person (maybe I am a crazy person, but I am also an engineer with a M.S. and years of experience with root cause analysis and building PCs) so maybe he does not take all his emails seriously. BUT - A 100% defect rate on all MSI x670e and x870e motherboards that MSI had failed to address for multiple generations now is a pretty damn big deal. Honestly, it impacts a lot more people than a lot of the other stories some of these YouTubers decide to cover. Maybe we need multiple people reaching out so they dont just ignore it in the flood of emails they might get.
How many users actually even know that their drives aren't working at the correct speed sometimes?
I didn't even realise until I benchmarked but how many even do that? Many users just switch on and use their machines - there's barely any difference between loading Windows at gen 1 or gen 5 PCI speeds, the same with games. The random read speeds are hardly any different across PCI generations
Also the problem is less likely to show up for users who power down every night. I only saw the problem when waking from sleep.
 
"Also the problem is less likely to show up for users who power down every night. I only saw the problem when waking from sleep."
This is false. The issue occurs on almost every boot. Consumer ignorance of the features they paid for is no reason for MSI to be able to sweep issues under the rug. These are $300+ motherboards (I'd be even more pissed if I paid $1000+ for the Godlike). If people are shelling out premium prices for borked garbage, I think it's warranted for companies to be publicly shamed into doing better by their customers. So far, official support channels are not delivering results.
 
"Also the problem is less likely to show up for users who power down every night. I only saw the problem when waking from sleep."
This is false. The issue occurs on almost every boot. Consumer ignorance of the features they paid for is no reason for MSI to be able to sweep issues under the rug. These are $300+ motherboards (I'd be even more pissed if I paid $1000+ for the Godlike). If people are shelling out premium prices for borked garbage, I think it's warranted for companies to be publicly shamed into doing better by their customers. So far, official support channels are not delivering results.
I totally agree about there needing to be more visibility on this. I kept my ticket active, will have to pester them again
I only ever saw the PCI problem on waking from sleep, never from a cold boot
 
"Also the problem is less likely to show up for users who power down every night. I only saw the problem when waking from sleep."
This is false. The issue occurs on almost every boot. Consumer ignorance of the features they paid for is no reason for MSI to be able to sweep issues under the rug. These are $300+ motherboards (I'd be even more pissed if I paid $1000+ for the Godlike). If people are shelling out premium prices for borked garbage, I think it's warranted for companies to be publicly shamed into doing better by their customers. So far, official support channels are not delivering results.
Agreed. The problem is 100% random and happens on any power cycle (restart, cold boot, wake from sleep). I can go five times in a row in any scenario, or in a stair-step fashion of x1, x2, x3, etc. Usually within five restarts though I can get it back to optimal speed if it did not land on it the first boot.

I sent a message to Gamer's Nexus today pointing out the problem. Maybe they will pickup on this if enough people complain. I also agree that many users here are probably unaware they even have a problem as not everyone is an enthusiast that likes to benchmark our systems like probably the majority on these forums.
 
Back
Top