Troubleshooting nvlddmkm errors with Throttlestop

[ROMAD]

TL;DR: mobo/gpu/cpu replaced, memory swapped, still have the issue. Looking to see what else might be causing it, using Throttlestop's logs as tool to help. Apologies for the long post. Skip to bottom for Throttlestop info.

I know this may not be the appropriate forum, but it does seem like Throttlestop might be able to help troubleshoot this. Laptop is a Sager NP8872T/Clevo PD70PNT, purchased in June 2022. Specs are as follows:

CPU: i7-12700H (no overclock)
GPU: RTX 3080ti 16gb (150w)
RAM: 32GB
Storage: 980 Pro 500gb (OS) & 980 Pro 2TB (data/games)

Background: I would get the following error randomly (system locks up for 10-30sec), but at least once per session when playing World of Warcraft (Classic). This is the ONLY game/app that causes this error.

The description for Event ID 0 from source nvlddmkm cannot be found. Either the component that raises this event is not installed on your local computer or the installation is corrupted. You can install or repair the component on the local computer.

If the event originated on another computer, the display information had to be saved with the event.

The following information was included with the event:

\Device\Video3
Error occurred on GPUID: 100

The message resource is present but the message was not found in the message table

Other games don't do this, though I don't play that many others (LostArk, Sims4, Warcraft III, Path of Exile and a few low-end Steam games). I would consider LostArk to be much more intensive than WoW, but I don't get the error in LostArk. Windows never crashes either.

Troubleshooting:

1. Laptop originally had 16gb of memory, I bought another 16gb, but this problem predates the memory upgrade. Though I did try memory in different slots and then one at a time, with no resolution.
2. I have upgraded/downgraded graphics drivers in search of a set that is better/worse than the last, but this seems to make no difference, so I stay updated. (Currently v528.02)
3. Used MSI Afterburner, Cinebench and some other software I cannot remember to try to duplicate, but still only occurs in WoW
4. Reinstalled Windows 10 from scratch (on advice from Sager) and installed all new drivers that were sent directly to me by Sager (no change)
5. Sent laptop in to Sager, where they were unable to duplicate with their tools/benchmarks, but replaced the motherboard, CPU & GPU anyway
6. Laptop originally had WD SN570 500gb SSD for the OS and HP EX920M for data/games, replaced both with above 980 Pro's
7. Reinstalled WoW from scratch on the new drive, completely redid the settings/addons
8. If I remove the games drive from this laptop and put it in my old one, I can play for hours and never get the error

I'll add more troubleshooting if I think of it.

Throttlestop:
I've tried to use Throttlestop in an effort to curb power usage, as I think that power might be the root cause of all this (no data to back that up yet). Here are my Throttlestop settings, as well as the log file that I got when the error last occurred (~11:39am EST).

 

[Twski]

Seems like it's GPU related.

1. Do you have an undervolt/overclock on the GPU? If so. try clearing that (since you tried almost everything I assume you'd already tried that, but just to be sure)
2. How many FPS do you run WoW? Since it's a very light game and you have a pretty powerful laptop, you may be rendering way too much FPS. I've heard cases of GPUs being unstable when running games in the thousands of FPS range. Did you ever try capping the FPS?
3. Do you get the error if you run it using the integrated GPU? Try it
4. Use the MSI Afterburner's logging in a few sessions. (Options > Monitoring > Log history to file). Then check the logs and look in the moment the error ocurred: is the frequency/voltage/power draw the same? Is it a spike?

 

[freeagent]

It sounds like it’s unstable because you are limiting power and as a result the nv driver is crashing.

 

[Mister300]

Try disabling any mem XMP/ AMD OC version profiles in your BIOS it fixed my errors got this resolved via Intel support.

 

[ROMAD]

1. Do you have an undervolt/overclock on the GPU? If so. try clearing that (since you tried almost everything I assume you'd already tried that, but just to be sure)
2. How many FPS do you run WoW? Since it's a very light game and you have a pretty powerful laptop, you may be rendering way too much FPS. I've heard cases of GPUs being unstable when running games in the thousands of FPS range. Did you ever try capping the FPS?
3. Do you get the error if you run it using the integrated GPU? Try it
4. Use the MSI Afterburner's logging in a few sessions. (Options > Monitoring > Log history to file). Then check the logs and look in the moment the error ocurred: is the frequency/voltage/power draw the same? Is it a spike?

1. I don't currently use any undervolt or overlock the CPU or GPU. Though I tried undervolting/clocking the GPU after problems first arose, but this was before the CPU, GPU & mobo got replaced.
2. I cap WoW at 165, because that's the limit of my monitor, but I'm open to try different settings, though I'm 90% sure I've dropped frames down to 60 and also tried setting WoW to use DirectX11 - Legacy, DirectX 11 and currently using DirectX12. Other games have the same cap and don't exhibit this behavior, but again, I'm open to try whatever.
3. I have not tried running the iGPU, but I'm curious what the upshot from that would be, seeing as the CPU, GPU and mobo have all been replaced, but I'll put that on the list. Edit: I ran with the iGPU and the game was stable, but obviously capped out at about 40fps in busy areas.
4. I'll see what MSI can log that HW Info hasn't given me. I didn't see anything conclusive from there, as far as dips in power, odd temp spikes, etc.

It sounds like it’s unstable because you are limiting power and as a result the nv driver is crashing.

I didn't start messing with power until after these problems started (pretty much right when I got the laptop). However, that's kinda why I'm here now. Mostly at my wits end with it. I have a suspicion my power brick may not be up to the task and it's the only thing that hasn't been replaced so far (it never gets "hot", only warm). They saddled this laptop with a 230W brick, which imo, is far too low. So if it's marginal at all, it could be causing problems.

Try disabling any mem XMP/ AMD OC version profiles in your BIOS it fixed my errors got this resolved via Intel support.

I'll double-check my BIOS, but the Sager/Clevo BIOS for this is mostly no-frills. (Just checked, there's no XMP settings)

 

[Twski]

Well, 165 fps shouldn't be a problem at all. The ocurrences I saw causing problems were in the >700fps range (in menus or extremely light games)>

Afterburner problably can't log anything extra, I didn't see that you already tried logging with HWinfo.

Yes, 230W is kind of low for you config. But it that was the problem, it would probably not happen when you limit the power on GPU/CPU

 

[ROMAD]

Ultimately, I'm just trying to gather more info. I'm not sure what would be the best way to gauge a laptop power supply's output in real time.

 

[Twski]

Ultimately, I'm just trying to gather more info. I'm not sure what would be the best way to gauge a laptop power supply's output in real time.

To gauge the output is a bit hard without tearing down the cable.

But you can gauge the input and then estimate the output.

To gauge the input, you may use a Kill-a-watt, a clamp current meter or (the cheap and dangerous way) a regular multimeter wired in series between the power cable and the wall socket

 

[ROMAD]

To gauge the input, you may use a Kill-a-watt, a clamp current meter or (the cheap and dangerous way) a regular multimeter wired in series between the power cable and the wall socket

I don't believe there's anything wrong with the wall socket, as this behavior happens wherever I plug the laptop in, whether it be at home or elsewhere. Lightbulb moment, no pun intended. I think I'm going to order a Kill-A-Watt.

Since I cannot update my original post (???), I'm updating here that the actual wattage of the GPU is listed as 155W (yikes). Hardly seems like there's much left over for the rest of the system, once you calculate what the CPU takes.

Here's a snippet from HWInfo while playing.

 

[Assimilator]

Hardly seems like there's much left over for the rest of the system, once you calculate what the CPU takes.

What did you expect, exactly, when you bought a device that crams components that can consume nearly 300W into a tiny chassis?

The problem is nothing to do with the hardware, it is how the hardware is configured. WoW is an ancient terribly-optimised game that uses massive amounts of CPU and very little GPU, and when your CPU - which is a power-guzzling pig of note - is maximally loaded, it draws so much wattage that the GPU momentarily doesn't have enough, which crashes the GPU and GPU driver, which is able to recover because the CPU is no longer drawing as much power at time it attempts recovery.

Probably the only way you can resolve this is by underclocking the CPU, GPU, or both. But really this is for Sager/Clevo to fix by issuing a BIOS update that slightly dials down the maximum power limits for these components so that they don't overdraw the system's total power budget when both are fully loaded.

 

[Vya Domus]

Power supply related perhaps ?

nvlddmkm errors can literally mean anything.