PCIe bug with M2_1 slot on X870E Tomahawk

[jlsmith.01712b902af]

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!

 

[rvillatoba14ec02d0]

I was thinking for a long time that it was something with my SSD, took me a while to notice that it was in pcie1 slot.
Samsung 990 PRO.
MSI B850 Edge Ti

BIOS version - 7E62v1A22

The issue is identical, sometimes it goes into pcie1 mode and speed drops to ~950mb instead of 7000+mb.
The issue is reproducible in about 1/3 of runs. Changes with reboot.

UPD: Now the issue is reproducible in most runs, not sure what is affecting it, but I wasn't able to get it into pcie4 for 5 times in a row.

Well that's weird. Because to my knowledge only X870E versions of MSI motherboards are affected by the M2_1 slot PCIe bug. No X870 is affected, nor the B850.
Which makes sense since the bug has existed since the previous generation, and only the X670E was affected (the X870E is a rebadged X670E). B650E (=X870) and lower versions were never affected by this bug. So what happens to you is extremely weird... I'd be tempted to say this is another bug specific to your card.

 

[b933930151502d6]

I'd be tempted to say this is another bug specific to your card.

...or just been lucky/not noticed it til now

 

[rvillatoba14ec02d0]

...or just been lucky/not noticed it til now

Have you heard of other B850 or X870 owners suffering from this bug? Because as far as I'm concerned, no, it's the first. If other owners of similar MSI boards have this problem, we can't find any trace of it on the forums, either here or on Reddit.

So yes, it's weird and it could be another bug that looks like the one of the X870E. Or maybe, as the poster says, it's the SSD.

To xxxdendelxx158802db, have you tried with another SSD? Just to be sure it's the same bug and rule out a possible defective drive? Does it happen only in the M2_1 port? What happens if you put the drive in another m.2 port? You said "Now the issue is reproducible in most runs", which means it wasn't the case before? Is it getting worse?

 

[leipp124c02a1]

Don't know if this is relevant, but:
There's an issue that affects some NVME M.2 devices in Linux. NVME spec has two low power modes "APST". Some devices get stuck in the low power mode.
The fix in Linux is to set the kernel boot parameter: nvme_core.default_ps_max_latency_us=0 which disables entering those lower power modes entirely.
When this issue happened to me I observed the device going from 4+GB/s down to 935MB/s.
Maybe there is a power save mode issue at hand here.

I have the MAG Tomahawk X870 (non E), and have a different problem:

While it always connects to my T705 at Gen 5.0x4 (32GT/s link speed), I'm only seeing 4.7GB/s reads. Meanwhile my older Gen 4.0x4 device in my old desktop sees a healthy 7.0GB/s (with 16GT/s link speed).
I would expect closer to 14GB/s. There are no errors logged and nothing to indicate why the speed is reduced. I've tested a few different kernels and they all see the same thing. So I'm guessing it's either an issue with the motherboard or the T705.

 

[Alan J T]

And why in the blue hell i need to configurate my Fans again ? Other boards left the Fan's profile. I'm with Arctic P12 Max, and it's really a pain going through that process again.

Was they Intel CPU boards

 

[C0K01_]

Have you heard of other B850 or X870 owners suffering from this bug? Because as far as I'm concerned, no, it's the first. If other owners of similar MSI boards have this problem, we can't find any trace of it on the forums, either here or on Reddit.

So yes, it's weird and it could be another bug that looks like the one of the X870E. Or maybe, as the poster says, it's the SSD.

To xxxdendelxx158802db, have you tried with another SSD? Just to be sure it's the same bug and rule out a possible defective drive? Does it happen only in the M2_1 port? What happens if you put the drive in another m.2 port? You said "Now the issue is reproducible in most runs", which means it wasn't the case before? Is it getting worse?

I don't have any other SSDs unfortunately, but I did try to put my 990 PRO into M2_4.
8 runs in M2_4:
PCIe4 - 8 times
PCIe1 - 0 times
=> Issue is not reproducible in M2_4.

In the M2_1 it's about 50% now, tried clearing CMOS - same.

The results are very random, to get a better % I need to run it more times, but it takes time.

So far in 13 runs in M2_1:
PCIe4 - 6 times,
PCIe1 - 7 times

Currently discussing it with MSI support. They are helpful so far

 

[C0K01_]

Don't know if this is relevant, but:
There's an issue that affects some NVME M.2 devices in Linux. NVME spec has two low power modes "APST". Some devices get stuck in the low power mode.
The fix in Linux is to set the kernel boot parameter: nvme_core.default_ps_max_latency_us=0 which disables entering those lower power modes entirely.
When this issue happened to me I observed the device going from 4+GB/s down to 935MB/s.
Maybe there is a power save mode issue at hand here.

I have the MAG Tomahawk X870 (non E), and have a different problem:

While it always connects to my T705 at Gen 5.0x4 (32GT/s link speed), I'm only seeing 4.7GB/s reads. Meanwhile my older Gen 4.0x4 device in my old desktop sees a healthy 7.0GB/s (with 16GT/s link speed).
I would expect closer to 14GB/s. There are no errors logged and nothing to indicate why the speed is reduced. I've tested a few different kernels and they all see the same thing. So I'm guessing it's either an issue with the motherboard or the T705.

Thank you for the info, haven't found something related in windows so far, but I tried running my system and SSD with:
- different params in Samsung magician (energy saving, standard, max perf);
- max power mode in windows without turning off drives setting (set to Never or high value);
- explicit Gen4 instead of Auto in M2_1 PCIe settings.
None of the above fixes the issue unfortunately.

 

[alexrose1uk]

Guys just to try and get some direction here; is it looking like the newest betas show substantial improvement towards fixing this speed issue for X870E based MSI motherboards; or is it looking like this is not going to go away, and I'm best to look at a board from another manufacturer?

I'd had my eyes set on the X870E Edge TI Wifi, but if its going to randomly collapse M.2_1 down to PCI-E Gen 1, and there are talks about MSI currently experiencing PCI-E training issues as well... well I'm rapidly starting to reconsider my second choice of board.

Unfortunately I want a board in silver/white, which limits my choices a lot, and the MSI had been my favoured choice due to the fact the PCI-E GPU slot never had it's bandwidth capped; I just had to decide if I want USB 40GBps or not. As it is, its starting to appear I'd be better taking that compromise elsewhere...

 

[jonomite]

Guys just to try and get some direction here; is it looking like the newest betas show substantial improvement towards fixing this speed issue for X870E based MSI motherboards; or is it looking like this is not going to go away, and I'm best to look at a board from another manufacturer?

I'd had my eyes set on the X870E Edge TI Wifi, but if its going to randomly collapse M.2_1 down to PCI-E Gen 1, and there are talks about MSI currently experiencing PCI-E training issues as well... well I'm rapidly starting to reconsider my second choice of board.

Unfortunately I want a board in silver/white, which limits my choices a lot, and the MSI had been my favoured choice due to the fact the PCI-E GPU slot never had it's bandwidth capped; I just had to decide if I want USB 40GBps or not. As it is, its starting to appear I'd be better taking that compromise elsewhere...

Unfortunately there are like 3 different sub-topics that have emerged in this thread, so it's gotten a little hard to follow. This topic of course is the original issue.

The latest BIOS for the X870E Tomahawk seems to resolve 99% of the issue with NVME drives reporting as PCIe 1 x 4. I and one other user reported what appears to be a very isolated instance when waking the CPU from standby and reporting a value other than expected (for me, it was PCIe 3 x 4 instead of PCIe 4 x 4). A reboot instantly fixed it and I have not noticed the issue again (and I check regularly).

I think there's a sub-topic in this thread addressing the same issue with the X870E Godlike. Users are reporting that MSI support is telling them it cannot be addressed through a BIOS update and to RMA the board. Folks are (understandably) very skeptical about that advice because it appears to be fixable through a BIOS update on the Tomahawk, and some users have reported that even after an RMA the issue persists.

I think a new sub-topic appeared recently regarding the issue in B850 boards. Not sure we've seen enough of that to be able to comment on whether it's a thing or not.

 

[igo154102da]

Weird, Godlike X870E here and the Crucial T705 1TB in M2_1 slot, and I'm yet to experience this drop to 900MB/s issue. I've been reading all threads where issues with regards to godlike boards are discussed and I've been checking my nvme drive speed regularly (both reported pcie gen and crystaldiskmark speed), but for some reason it doesnt happen to me. Not that I'm complaining, but just to inform everyone here that the probability of this bug isn't 100%. However, I'm sure that at least one person with this bug has RMAd his/her board to MSI to reproduce (duplicate?) this issue and I'm sure they have. Likely, it's not on their priority list, just like the DD3 firmware and bugs around it.

 

[crai151502d0]

Weird, Godlike X870E here and the Crucial T705 1TB in M2_1 slot, and I'm yet to experience this drop to 900MB/s issue. I've been reading all threads where issues with regards to godlike boards are discussed and I've been checking my nvme drive speed regularly (both reported pcie gen and crystaldiskmark speed), but for some reason it doesnt happen to me. Not that I'm complaining, but just to inform everyone here that the probability of this bug isn't 100%. However, I'm sure that at least one person with this bug has RMAd his/her board to MSI to reproduce (duplicate?) this issue and I'm sure they have. Likely, it's not on their priority list, just like the DD3 firmware and bugs around it.

Are you also running a Gen5 in M2_2? In addition if you do, do you have the BIOS setting ASM4242/M.2_2 switch set to M.2_2x4?
Seems like all of us having this problem are trying to utilize two Gen5 Nvme drives at their full speed, if I'm not mistaken.

 

[kolkoo]

Are you also running a Gen5 in M2_2? In addition if you do, do you have the BIOS setting ASM4242/M.2_2 switch set to M.2_2x4?
Seems like all of us having this problem are trying to utilize two Gen5 Nvme drives at their full speed, if I'm not mistaken.

Nope. I just built my system two weeks ago and found out last weekend I have the issues with 4TB Crucial T705 in PCIe 5.0 M2_1 and PCIe 4.0 990 Pro in M2_2. When building my system I had already flashed all bios to latest available. I moved M2_1 -> M2_2 and M2_2 to M2_3, disabled USB4 and I'm getting full speeds.
I just ordered a "Samsung 9100 PRO NVMe M.2 SSD, 4TB, PCIe 5.0" to put in M2_1 slot as well to check speeds or fixes without disassembling and having extra slow storage if they never fix it.

 

[crai151502d0]

Nope. I just built my system two weeks ago and found out last weekend I have the issues with 4TB Crucial T705 in PCIe 5.0 M2_1 and PCIe 4.0 990 Pro in M2_2. When building my system I had already flashed all bios to latest available. I moved M2_1 -> M2_2 and M2_2 to M2_3, disabled USB4 and I'm getting full speeds.
I just ordered a "Samsung 9100 PRO NVMe M.2 SSD, 4TB, PCIe 5.0" to put in M2_1 slot as well to check speeds or fixes without disassembling and having extra slow storage if they never fix it.

So yes, the problem is still there. Many of us that bought this board did so for two Gen5 slots as it is advertised. Right now it seems it's just one truly working Gen5 at the expense of USB lanes. M2_1 is still bufu'd

 

[igo154102da]

Are you also running a Gen5 in M2_2? In addition if you do, do you have the BIOS setting ASM4242/M.2_2 switch set to M.2_2x4?
Seems like all of us having this problem are trying to utilize two Gen5 Nvme drives at their full speed, if I'm not mistaken.

Yep, I have another Crucial, but T700 4TB there:

And this is what I get every time/reboot on M2_1:

 

[igo154102da]

So yes, the problem is still there. Many of us that bought this board did so for two Gen5 slots as it is advertised. Right now it seems it's just one truly working Gen5 at the expense of USB lanes. M2_1 is still bufu'd

I am not denying the problem, just saying it's not happening to 100% of Godlike X870E users. I've been on this board since Jan 2025, keeping an eye on the SSD performance in that slot, but havent yet seen it switching to PCIe 1 x 4

 

[b933930151502d6]

I am not denying the problem, just saying it's not happening to 100% of Godlike X870E users. I've been on this board since Jan 2025, keeping an eye on the SSD performance in that slot, but havent yet seen it switching to PCIe 1 x 4

What's the manufacture date of your board? It's in the serial number, like mine is 2412 so December 2024

 

[igo154102da]

What's the manufacture date of your board? It's in the serial number, like mine is 2412 so December 2024

2410, probably October then.

 

[crai151502d0]

What's the manufacture date of your board? It's in the serial number, like mine is 2412 so December 2024

Mine is the same 2412, and has the issue.

 

[1160b02f1]

Yep, I have another Crucial, but T700 4TB there:

View attachment 201921
And this is what I get every time/reboot on M2_1:

View attachment 201922

So I have been trying to find some way of tracking hardware revisions to see if it is a physical issue and MSI actually solves it, how would we know which boards have the updated hardware? They sure as hell do not seem to be transparent enough to actually tell us. And I think I found something through a PowerShell command that the documentation defines as reporting out a mobo's "Version of the physical element" "define any component of a system that has a distinct physical identity." so this might tell us the hardware revision? I standby to be corrected if anyone knows more than me here.

Anyway, Can you tell us what this tells you under "Version" if you enter this into Powershell? -->
gwmi win32_baseboard | FL Product,Manufacturer,SerialNumber,Version

Mine reads "Version 1.0" and I have the problem. I wanted to use this on more recently manufactured boards to see if that number ever changes to a 2.0, But since you think your board works, you have me curious what yours reports for that too.

Also, could you please tell us what BIOS version you are using and if you changed any settings?

 

[crai151502d0]

So I have been trying to find some way of tracking hardware revisions to see if it is a physical issue and MSI actually solves it, how would we know which boards have the updated hardware? They sure as hell do not seem to be transparent enough to actually tell us. And I think I found something through a PowerShell command that the documentation defines as reporting out a mobo's "Version of the physical element" "define any component of a system that has a distinct physical identity." so this might tell us the hardware revision? I standby to be corrected if anyone knows more than me here.

Anyway, Can you tell us what this tells you under "Version" if you enter this into Powershell? -->
gwmi win32_baseboard | FL Product,Manufacturer,SerialNumber,Version

Mine reads "Version 1.0" and I have the problem. I wanted to use this on more recently manufactured boards to see if that number ever changes to a 2.0, But since you think your board works, you have me curious what yours reports for that too.

Also, could you please tell us what BIOS version you are using and if you changed any settings?

Well, this brings up another interesting issue. The reported Serial Number does not remotely even match what's on my MB box. Check against yours please.