Even after replacing the CPU, the PCIe slot version is still stuck.

[gjrtm343154402df]

My previous setup was:
9700X
X870 Tomahawk WiFi
G.Skill 64GB CL30 6000MHz
RTX 5080
FSP 1000W PSU
Everything was working perfectly fine. But after replacing my CPU with the 9800X3D, everything went completely wrong.

From the very first boot with the 9800X3D, I immediately noticed stuttering on the Windows lock screen—my mouse, keyboard, and display were lagging. It felt like the system was freezing every 3–5 seconds. As the issue worsened, the screen would go black, or I would get a Video Scheduler Internal Error blue screen, causing the system to restart. However, sometimes after a reboot, the system would work normally.

When the system was stuttering, I checked GPU-Z, and I noticed that the PCIe version was stuck at 1.1x16. Even under load, it remained locked at 1.1x16. In some cases, it was stuck at 2.0x16 or 3.0x16, causing severe lag. During these moments of lag, even when the system was idle, the GPU usage would spike abnormally. And when the system did work "normally," the PCIe version remained locked at 5.0x16, rather than dynamically adjusting between 2.0x16 (idle) and 5.0x16 (load) as it should.

Since all of these problems started only after installing the 9800X3D, I decided to swap back to my 9700X, and everything returned to normal. Even in GPU-Z, I could see that PCIe behavior was back to normal—2.0x16 in idle and 5.0x16 under load.

The 9800X3D has already been sent for RMA and replaced. and I have installed the latest version, 1A26 Beta, but the issue remains the same.

 

[EsSeNce25]

Reportedly, a bios update came out (possible) addressing this exact issue but a few days ago. So check for that as well.

Yes the one with: - "Improved PCIe-based devices compatibility under certain CPU or system combinations."
On my MSI X870 Tomahawk v1.1 the issue still persists with that new bios. I did notice it happening less often.
With the solution I posted above, the problem is gone completely.
It is something that Microsoft will have to fix as it is related to the graphics overlay system.

Everyone testing this, let me know your findings.

 

[craigcoope151302d0]

Yes the one with: - "Improved PCIe-based devices compatibility under certain CPU or system combinations."
On my MSI X870 Tomahawk v1.1 the issue still persists with that new bios. I did notice it happening less often.
With the solution I posted above, the problem is gone completely.
It is something that Microsoft will have to fix as it is related to the graphics overlay system.

Everyone testing this, let me know your findings.

From what I have read if you do this fix it causes terrible stutter if you are using frame generation. Seeing as I game at 4k, that is an issue.

 

[Mark1234567]

What helped me was raising VSOC voltage instead of lowering it (1,27v). Almost every boot PCIe is at 5.0x16 but locked.
And now I changed spread spectrum from auto to disabled instead of enable and now PCIe also changes between 2.0-5.0 correctly.
This is with MPG X670E CARBON WIFI and with AGESA PI 1.2.0.3g.

 

[EasterEEL]

From what I have read if you do this fix it causes terrible stutter if you are using frame generation. Seeing as I game at 4k, that is an issue.

I occasionally (a couple of times a month) get a black screen with nvlddmkm 153 events when browsing the internet using Edge or Firefox. Always recovers after a couple of secs. Never had a problem playing games. I'm going to spend a bit of time looking at this and have disabled MPO / multiplane overlay to see if fixes. If it does, will then look at possible stuttering in games using Frame Generation, although I have read that it can fix stuttering not make it worse. Do you have source so I can read up?

 

[Jarous1]

I have exactly the same symptoms. Almost every time I start or restart my PC, the PCIe runs at a different speed. Even I’ve forced it to 5.0 in the BIOS, the computer completely ignores it and sets it however it wants. At 1.1, of course, there’s stuttering until the system freezes and restarts.
It’s surprising that MSI still hasn’t fixed this after a whole year

 

[craigcoope151302d0]

I think I finally found out what is causing this. The culprit is MPO / multiplane overlay used in Windows. To test for this do the following:

Using Edge/Chrome/UngoogledChrome, go to yandex and do a random search then hit F11 for full screen, then run the mouse pointer over all the results in circular motion touching the top of the screen on every rotation, keep doing it over and over. Have Latency Monitor tool open and check it regularly. You will notice super high DPC latency spikes. What happens the whole system stutters. (audio will stutter if you have something playing, USB sound devices will also stutter since this is system wide freeze)

The same happens when you bring the PC out of sleep or at boot. Randomly, the Multiplane Overlay used by modern windows, causes the PC to briefly freeze stutter therefore negotiating a slow GPU link speed. This can happen at boot or when waking up PC from sleep.

To fix this, download MPOGPUFIX v6.6 tool and flip the MPO Fix toggle switch to enable it. Reboot and if I'm correct, you will never see this link speed problem again. Please test this and report back. So far on my end the problem no longer occurs after using the tool.
The tool is not intrusive it just adds a registry entry to disable MPO and when switched off it removes it.

(IF my finding is correct this is a huge discovery as many people are affected by this)

I'll try to find the link. I searched up MPOGPUFIX after you posted about it. Some users reported it made stutter worse if you use frame generation.

 

[Mika the second one]

This seems to be an MSI BIOS issue.
My 9950X3D paired with a 9070 XT on an X670E Carbon always showed only 16 x 2.0 and 16 x 1.1 in GPU-Z, and if I restarted the PC enough times, it would eventually show 16 x 5.0. But that would disappear again after the next restart.
The only thing that helped me was setting the PCIe lanes in the BIOS to 8x8. Then I at least always had 16 x 4.0

By the way, the BIOS was always up-to-date.

But this limitation bothered me. I’ve now bought a motherboard from a different manufacturer, and everything works as it should.

 

[craigcoope151302d0]

This seems to be an MSI BIOS issue.
My 9950X3D paired with a 9070 XT on an X670E Carbon always showed only 16 x 2.0 and 16 x 1.1 in GPU-Z, and if I restarted the PC enough times, it would eventually show 16 x 5.0. But that would disappear again after the next restart.
The only thing that helped me was setting the PCIe lanes in the BIOS to 8x8. Then I at least always had 16 x 4.0

By the way, the BIOS was always up-to-date.

But this limitation bothered me. I’ve now bought a motherboard from a different manufacturer, and everything works as it should.

From what i've read it also happens on some Asus and Gigabyte setups. It does seem to be much more prevalent on MSI motherboards though.

 

[sblantipodi]

My previous setup was:
9700X
X870 Tomahawk WiFi
G.Skill 64GB CL30 6000MHz
RTX 5080
FSP 1000W PSU
Everything was working perfectly fine. But after replacing my CPU with the 9800X3D, everything went completely wrong.

From the very first boot with the 9800X3D, I immediately noticed stuttering on the Windows lock screen—my mouse, keyboard, and display were lagging. It felt like the system was freezing every 3–5 seconds. As the issue worsened, the screen would go black, or I would get a Video Scheduler Internal Error blue screen, causing the system to restart. However, sometimes after a reboot, the system would work normally.

When the system was stuttering, I checked GPU-Z, and I noticed that the PCIe version was stuck at 1.1x16. Even under load, it remained locked at 1.1x16. In some cases, it was stuck at 2.0x16 or 3.0x16, causing severe lag. During these moments of lag, even when the system was idle, the GPU usage would spike abnormally. And when the system did work "normally," the PCIe version remained locked at 5.0x16, rather than dynamically adjusting between 2.0x16 (idle) and 5.0x16 (load) as it should.

Since all of these problems started only after installing the 9800X3D, I decided to swap back to my 9700X, and everything returned to normal. Even in GPU-Z, I could see that PCIe behavior was back to normal—2.0x16 in idle and 5.0x16 under load.

The 9800X3D has already been sent for RMA and replaced. and I have installed the latest version, 1A26 Beta, but the issue remains the same.

MSI mobos are crap, same issue here on X670E ACE on a 9950X3D with the Agesa 1.3.x BIOS.
With Agesa 1.2.x no issue.

I'm sure that the issue was introduced in the latest BIOSes with Agesa 1.3.x because all BIOSes with Agesa 1.2.x works perfectly fine for months.

 

[joeyale15d902ea]

I have to wonder if this is an AM5 issue…I just switched from a Gigabyte X870e Aorus Master to the X870e carbon because of massive pcie issues with my 5080. I would get hundreds if not thousands of pcie lane errors (actual errors, not the recovery counter) and frequent 116/Video TDR blue screens. So far the new MSI board has been flawless. I tried everything with the Gigabyte board, including 3 different 5080s, and the only thing that made it somewhat stable was locking the GPU to maximum performance.

I’m not sure if I’m just lucky or if the latest bios fixed it or what, but I’d say this isn’t only an MSI issue.

 

[Notwalkin]

I have to wonder if this is an AM5 issue…I just switched from a Gigabyte X870e Aorus Master to the X870e carbon because of massive pcie issues with my 5080. I would get hundreds if not thousands of pcie lane errors (actual errors, not the recovery counter) and frequent 116/Video TDR blue screens. So far the new MSI board has been flawless. I tried everything with the Gigabyte board, including 3 different 5080s, and the only thing that made it somewhat stable was locking the GPU to maximum performance.

I’m not sure if I’m just lucky or if the latest bios fixed it or what, but I’d say this isn’t only an MSI issue.

So, something I've been dealing with among many other things...

Naks sent count in hwinfo64 - i get these in games and can replicate them pretty easily. Not a big deal but others say they don't get them at all.

Have tried 2x 5090, 1x 5070 ti, several B650 and x870e motherboards, 7600x, 9800x3d and 9850x3d, Corsair 1200w psu and Seasonic 1600w psu.
Quite literally, everything but the Platform (am5) and house, have been swapped.

Cyberpunk benchmark, quite often will cause 1-2 naks sent mostly from loading in, some motherboards seem to not cause them.
Hades 2 main menu, sitting on main menu will reliably generate them.

I can "fix" them by disabling CPU ASPM in bios + Forcing prefer maximum performance in Nvidia app/control panel.
Again, some motherboards behave better w/o disabling ASPM, i would guess it's BIOS differences on what ASPM on auto actually is doing.

I do think Am5 chipset is wobbly at best.

 

[Deleted member 1047288]

I fixed this issue on my Carbon x870e running A 9800x3d & 5080 by using bios 1A10.
I spent days trouble shooting and got quite irritated.
If I want to use any newer bios with any stability I have to Turn on Spread Spectrum & also turn off ASPM in bios.
I also had a 1% lows stuttering issue while gaming I fixed by turning on C-States in bios.

 

[vic123159d02e5]

Wrong thread sorry

 

[craigcoope151302d0]

Yes the one with: - "Improved PCIe-based devices compatibility under certain CPU or system combinations."
On my MSI X870 Tomahawk v1.1 the issue still persists with that new bios. I did notice it happening less often.
With the solution I posted above, the problem is gone completely.
It is something that Microsoft will have to fix as it is related to the graphics overlay system.

Everyone testing this, let me know your findings.

I decided to try out your fix. Unfortunately it did not fix the stuck on lower pci-e speed. First boot after fix was 2.0x16, nest was 3.0x16. Sorry to be the bearer of bad news.