X870 Tomahawk Cold Boot Issues - How many of us are there?

J

jonjon12315c602ec

New member
Joined
Oct 4, 2025
Messages
4
I've posted before about this issue in other threads and other subreddits, and from what I can gather this is specifically a problem only on MSI X870/X870e boards when paired with an X3D processor.

Basically, when the system has been shut down for some time, on start up the sound lags, the screen becomes laggy/jittery and 50% of the time the system crashes if you don't restart. From checking various threads it seems that it's an issue that was introduced in a BIOS revision where the main PCIe slot is forced to work at x1 speed. This happens at completely random times, it can be fine for days and then happen twice in a row.

I did so much troubleshooting before realizing it was a BIOS problem, I changed GPU, PSU, M.2 drives and upgraded from a 7800X3D to a 9800X3D along the way with the issue persisting. I've tried different BIOS configurations, and a fresh windows install didn't help either.

Before realizing it was a BIOS problem I contacted MSI and told them about the issue, where they offered an RMA which I didn't take, as I was worried the problem would just persist on another board. MSI are now trying to say it's either a GPU problem, or suggesting it may be due to having an M.2 occupying the M.2_1 slot which makes no sense whatsoever, nothing should share bandwidth with the main PCIe1 slot as far as I'm aware.

I'm trying to tell MSI that it's clearly a BIOS issue and they need to look into a firmware update rather than an RMA, but they seem to be skirting around accepting this or even entertaining the idea.

Here are some other threads I quickly found detailing the same or similar issue;




And my personal experience;


If you have experienced this, and can tell me what the last stable BIOS was, I can feed this back to MSI and push for a fix. I also suggest you log a ticket too, the more eyes we have in it the better the chances for a fix are.
 
Sort by date Sort by votes
So how long is "some time" , is that overnight? Or 24H minimum?
Because, I still have the usb quirks when booting, where sometimes the keyboard isn't activated in time to get into the bios, cold boot much faster than soft reboot (allthough a soft reboot is pretty decent now) still a duplicate boot drive in bbs priorities section, where the second is my primary os and the first my dual boot. (no bcdedit nor bcbootrec/diskpart command seems to fix it either)
But for the rest it's pretty good.
 
Upvote 0 Downvote
tfbisqui156f02d4 said:
So how long is "some time" , is that overnight? Or 24H minimum?
Because, I still have the usb quirks when booting, where sometimes the keyboard isn't activated in time to get into the bios, cold boot much faster than soft reboot (allthough a soft reboot is pretty decent now) still a duplicate boot drive in bbs priorities section, where the second is my primary os and the first my dual boot. (no bcdedit nor bcbootrec/diskpart command seems to fix it either)
But for the rest it's pretty good.
Click to expand...
The time frame can be completely random, the only consistency is the fact it only happens from a cold boot. I turn my OC off every night, some mornings it can be fine, others it will stutter and require a restart.

Fortunately I've not had any USB issues.
 
Upvote 0 Downvote
If you're collecting data and responses, you should also ask whether they are booting from an NVME M.2 and which slot that drive is in. The amount of NVME M.2 issue posts on a weekly basis is insane. I use only SATA SSDs, and have never had an issue on any build. I'm currently using an older BIOS the board came with (no need to update right now).
 
Upvote 0 Downvote
I'm effected by this, but manually setting the PCIe lane modes helps.
Having said that, I also have issues with voltages (VDDIO and VDDP) needing to be manually set to use expo, and wrong cores prioritised on 7950x3d when using frequency offsets.
I'm feeling like these bios's are rushed a bit.
 
Upvote 0 Downvote
I have a very similar issue with my X870E Carbon WiFi (Ryzen 7 9800X3D + RTX 5080) — after updating to the 7E49v1A70 BIOS I started getting cold boot freezes as well. At first thought it was my M2 drive but I have it cloned and same thing started occuring on the other one.
Thing is, my PCIe settings were already set to the correct standard: 5.0 (one M.2 + GPU) and 4.0 (two M.2). I even changed the battery on the motherboard, which is a pain, but it did not fix the issue.

Eventually, after talking a bit with the AI, the AI helped me create a registry key entry that disabled the NVMe power saving state (APST), which can cause issues.
Surprisingly THAT — along with disabling Fast Boot — resolved the cold boot issue completely for me, however… every now and then I still get freezes on shutdown every now and than (like 1-2 a week, espetially after longer gaming sessions), which did not occur in the past, before the BIOS + Windows 11 25H2 update.

Honestly, it feels like those BIOS updates are a bit rushed. They sometimes appear two weeks after another and may not properly tested.
I hope MSI reps read these forums, because along with Reddit threads there is an issue there, and it’s not on us to fix.

Running RAM at stock, PBO off, etc. does not resolve these issues, nor do commands like DISM or SFC in Windows — just in case someone comes with troubleshooting ideas like that.
All the drivers, OS etc are up 2 date. I also did perform CMOS reset etc. nothing.
 
Upvote 0 Downvote
Enabling/disabling spread spectrum seems to have solved it for me (I can't remember which one, it was the option other than what is set by default). I will say however, that MSI have been pretty unhelpful. The only resolution they have offered is to RMA the board, when a quick 30 minutes online will show that this seems to be a BIOS issue from many reports online. I will continue to keep an eye on it though.
 
Upvote 0 Downvote
Too many of us having the same issue to RMA I would say. Also I had exactly the same issue on a differem mobo - MSI X870E Carbon Wifi -and fixed it by disabling NVME power saving state (APST) in the registry. My RIG just suffer sporadical freezes on shutdown now :( and all was 100% working without the recent BIOS updates (said updates as they released way to many for me to narrow it down in a short period of time). RMA takes 4-6 weeks - and that's the last resort - which I this case may not be even a solution.
 
Last edited:
Upvote 0 Downvote
jonjon12315c602ec said:
The time frame can be completely random, the only consistency is the fact it only happens from a cold boot. I turn my OC off every night, some mornings it can be fine, others it will stutter and require a restart.

Fortunately I've not had any USB issues.
Click to expand...

This is exactly what happens to me and it's driving me crazy, because it's completely random. And it's true, it always happens on a cold start, never on a warm restart, so to speak. And it's not that it always happens when I turn on the PC after several hours of being turned off, but the stuttering is more common than not.

I'm using BIOS 1A69 and I can't remember which was the last stable BIOS in this regard.
 
Upvote 0 Downvote
I've noticed that when I boot into a stutter environment with my MSI x870 Tomahawk, the event viewer contains entries like:

Code: 
The driver \Driver\WUDFRd failed to load.
Error setting traits on Provider
The luafv service failed to start due to the following error: This driver has been blocked from loading

From wading through all the guff and theories trying to solve this issue, trying a lot of ideas. I now believe this is caused by a race condition. Loading things in the wrong order or too quickly or moving on before something is ready... something like that. My system is fully up to date and I'm on the latest BIOS A69.

What actually helped the most was disabling the memory integrity under `Windows Security > Device Security > Core Isolation > Memory Integrity` whilst waiting for a bios/agesa update.
It's not 100% but it's reduced the stutter lottery to about 1 in 30 boots for me (from 1 in 2).
 
Upvote 0 Downvote
jonjon12315c602ec said:
Enabling/disabling spread spectrum seems to have solved it for me (I can't remember which one, it was the option other than what is set by default). I will say however, that MSI have been pretty unhelpful. The only resolution they have offered is to RMA the board, when a quick 30 minutes online will show that this seems to be a BIOS issue from many reports online. I will continue to keep an eye on it though.
Click to expand...
Using spread spectrum will not fix this permanently, it will only reduce the probability to get a bad boot, but the problem will eventually return. It can boot fine a month, but then it doesn't.
I had this problem for 6 months, and it only ended after replacing the motherboard for the same model (X870 Tomahawk Wifi). MSI is right to tell you to RMA it.
I have the latest BIOS now, no problems (yet) and no spread spectrum or voltage changes needed to get it boot with PCIe 5. If you need spread spectrum enabled, that means your board is somehow defective. Spread spectrum only masks the real problem, but doesn't repair it.

I'm just here to tell you to not get insane about this problem by trying to find solutions when there are none – replacing is really the only option, even if it sucks, believe me. It will never get fixed, because it's NOT a bios fault. It's a defective board (and obviously there are many of them, but not only from MSI).
 
Upvote 0 Downvote
Exact same issue here.

Pre-build Powered by MSI computer.

MSI X870 GAMING PLUS WIFI (latest 7E51v1A69 bios)
NVIDIA MSI GeForce RTX 5080 16G SHADOW 3X OC
AMD Ryzen 7 9800X3D - 8 Cores
32GB DDR5-5600 Kingston Fury Beast RGB (KF556C36-16)

I can confirm the issue likely lies with the GPU unable to negotiate the full PCIe 5.0, but instead getting stuck with a much lower PCIe Gen.

- On a good boot the GPU can run at PCIe 5.0, and all seems well, stable, and fast..
- If the GPU boots to a maximum of PCIe 3.0 or 4.0 instead, you won't really notice issues (on the desktop) but obviously it's not good and comes with a heavy performance penalty in case of PCIe 3.0.
- If the GPU boots to an even lower Gen (to PCIe 1.1 or 2.0) you'll be greeted by a Windows stuttery hell, mouse skips, screen freezes and reboots with black screen instead of a BSOD as a result.
- 'Event Viewer' and Minidump analyze tools like 'WhoCrashed' and 'BlueSCreenView' will point to the GPU drivers like nvlddmkm.sys, dxgkrnl.sys and watchdog.sys with descriptions like; " VIDEO_TDR_ERROR", "VIDEO_ENGINE_TIMEOUT_DETECTED" and "VIDEO_SCHEDULER_INTERNAL_ERROR", likely because the latency went through the roof.

You might also see some graphic artifacting as VRAM buffers overflow.

You can check which PCIe Gen version your GPU is stuck running at from boot with the program TechPowerUP GPU-Z.
* Note that in order to get a good measurement you need to use the "PCIe-Express Render Test" (click on the question mark -> Start Render Test) to be sure your GPU switches to the maximum PCIe Gen is was negotiated at boot as throttling PCIe to a lower Gen (even down to 1.1) is usually normal and healthy when the GPU is not under any load,

Look to the right of 'Bus Interface' where it will note something like "PCIe x16 5.0 @ x16 3.0"* or "PCIe x16 5.0 @ x16 4.0" which means that, though the GPU is capable of PCIe 5.0, the GPU is stuck @ a maximum of PCIe 3.0.

BTW: Putting the computer to sleep and waking it again up might boost the GPU from PCIe 3.0 to 4.0 or back to PCIe 5.0, but it might also wake up with a black screen and reboot instead (I assume because it then woke up to PCIe 1.1 or perhaps 2.0)
 
Last edited:
Upvote 0 Downvote
reijkelho15d702ea said:
Exact same issue here.

I can confirm the issue likely lies with the GPU not booting to PCIe 5.0, but a much lower version.

On a good boot the GPU goes to PCIe 5.0 and all seems well.
If the GPU boots to PCIe 3.1 or 4.0, you won't really notice issues (on the desktop), but obviously it's not good.
If the GPU boots to an even lower version than that (to PCIe 1.1 or 2.0) you'll be greeted by stutter hell, freezes and reboots with black screen (No BSOD). And Event Viewer and tools like WhoCrashed will point to the GPU drivers, like nvlddmkm.sys, dxgkrnl.sys, and watchdog.sys which will have crashed because the latecy goes through the roof.
You might also see some graphic artifacting as VRAM buffers overflow.

You can check which PCIe your GPU is stuck running at with the program TechPowerUP GPU-Z.
Look at Bus Interface where it will say something like "PCIe x16 5.0 @ x16 PCIe 3.0"* which means that, though the GPU is capable of PCIe 5.0, it's actually stuck running @ maximum of PCIe 3.0.
* Note that you need to use the render test under the question mark to put your GPU at max PCIe as throttling PCIe to a lower version is normal when not under any load.

BTW: Putting the computer to sleep and waking it again up might boost the GPU from PCIe 3.0 or 4.0 back to PCIe 5.0, but it might also wake up with a black screen and reboot instead (I assume because it then woke up to PCIe 1.1 or perhaps 2.0)

MSI X870 GAMING PLUS WIFI (latest bios)
NVIDIA MSI GeForce RTX 5080 16G SHADOW 3X OC
AMD Ryzen 7 9800X3D - 8 Cores
32GB DDR5-5600 Kingston Fury Beast RGB (KF556C36-16)
Click to expand...
Exact same issue with my setup as well.

I have just recently finished my new PC Build 2-3 Weeks ago. At first there weren't any problems at all. Until i updated the BIOS to the latest version, then this exact problem started occuring almost every time. Whenever i come home from work, the PC is "Cold" and booting almost always fails. I have to manually power it off, because it gets stuck, and usually the second boot is succesfull. Today, however, i tried 4 boots and all of them failed. It always got stuck and i got really mad. I have 3x DP Monitors + 1x HDMI connected. I disconnected the HDMI and noticed it was really hot, despite the actual screen being disabled from the OS. After disconnecting the HDMI, the PC booted normally. It may indeed be an issue with the PCIe 3.0, since i am still using my old 1080Ti, which does not support PCIe 5.0. I tried all the fixed i saw. PCIe is forced to 3.0 from the BIOS settings. They should fix it, since it is a real issue, causing headaches.

MB: MSI Pro X870E-P (7E70v2A31 / 2025-09-24)
CPU: AMD Ryzen™ 7 7800X3D 8C 16T
CPU COOLER: Corsair iCUE H100i Elite Capellix
GPU: NVIDIA GEFORCE GTX 1080Ti FE 11GB GDDR5X
RAM: Corsair Vengeance 2x16GB DDR5 6000MHz
SSD: Kingston NVMe 1TB
SDD2: Kingston NVMe 1TB
HDD: Seagate IronWolf Pro 4TB
CASE: Corsair 4000D Airflow
PSU: Corsair AX850 80+ Gold
OS: Windows 11 Pro 64-bit
DISPLAY 1: 27" FHD IPS LG 144Hz G-Sync
DISPLAY 2: 27" FHD VA LG 165Hz Free-Sync
DISPLAY 3: 27" FHD VA LG 165Hz Free-Sync
DISPLAY 4: 43" 4K LG TV
 
Upvote 0 Downvote
These stuttering, freezing crashes on boot-ups seems to have been a well enough documented issue since the Bios update after November 2024, though not likely to be acknowledged by MSI, much less fixed, anytime soon.

Alternative conclusions would be either:

- The GPU might not be seated correctly in the PCI slot or have bad contacts, causing the connection lanes to be noisy and negotiate to a lower Gen PCIe.
- Thermal expansion on lanes as temperatures increase are also at play. Because of the cold boot factor.
- Halfway Defective GPU or motherboard lanes.

The motherboard is a cheapy version with cheapy lanes. Issues happen.

Regardless, it's probably best for me to RMA this computer while I still can.
 
Last edited:
Upvote 0 Downvote
Back
Top