Laptop GPU lag spikes due to power supply issues? (PerfCap reason PWR)

[Mark_Morgan]

System: Asus zenbook pro duo UX582LR
CPU: Intel Core i7-10870H CPU @ 2.20GHz 2.21 GHz
GPU: GeForce RTX™ 3070 Laptop GPU
RAM: 32GB

Ok, so for quite some time I've been having issues with my Zenbook Pro Duo when the GPU is at maximum utilization and it's driving me a bit crazy, as it doesn't seem to be a common issue with this specific machine.
After googling around and trying different things (re-installing Windows, drivers, undervolting) without success I hope that I finally got a lead by doing analysis with GPU-Z (great software btw ).
At first, I thought the issue was just due to thermal throttling, it is a non-gaming laptop after all, but after finding no one else with this specific problem on this specific laptop I guess the issue is something more specific and honestly, I'm puzzled.

So the problem is as follows: Whenever I play a demanding game, after a few minutes of smooth gameplay, I start to get very heavy lag spikes every couple of seconds making the game unplayable. Looking into the task manager I can see that every time a lag spike occurs my GPU utilization drops from 100% to 0% for a second. Now looking at GPU-Z sensors during gameplay, every time this happens the PerfCap Reason switches from Thrm to Pwr for a split second. Now this and the fact that during the first couple minutes of smooth gameplay the GPU temp is just as high as when it starts lagging kinda made me rethink my previous assumption that the issue was coming from overheating and consider a power issue instead, but I'm honestly not sure anymore. Any Ideas?

Attached is an annotated GPU-Z screenshot, I also un and replugged my power supply during playing which for some reason made the game run smoothly again for a bit.

 

[jon_joy_1999]

I think you've got cooling issues in your laptop. You can see the temp rise, and when it peaks the GPU clock drops until it goes into lag city. This is just speculation on my part, but I'm guessing the throttle reaches a point where it just turns something off, and then back on and the card plays catchup

 

[WingCommanderBob]

Some throttling is causing a momentary drop in temps, so GPU boost provides more power, at which point it runs into the power limit and then a shortly after runs into the thermal limit again. The throttling behaviour could be being caused by prolonged high GPU chip temps (though imo that's unlikely), or may also be something for which the telemetry is not exposed to the OS such as vrm or vram temps. Personally I'd take a look under the heatsink for anything obvious like dried out thermal pads and take the opportunity to repaste with something better than whatever Asus used.

 

[jon_joy_1999]

Before you disconnect the heatsink, I'd clean the radiator fans and fins, and if that doesn't fix the issue, go forward with Bob's suggestions.

 

[eidairaman1]

Todays laptops are not laptops at all but jokebooks