I'm trying to troubleshoot a performance problem that began about five days ago on my gaming laptop. It has an Intel Core i7-12700H, an RTX 3060, 32 GB of DDR4 RAM, and the original power adapter. During gaming, the CPU reaches around 90–91°C and the GPU reaches approximately 86–87°C.
After roughly 30 minutes, the GPU clock drops to about 900 MHz. The game can remain relatively smooth while the clock stays there, but whenever it briefly tries to boost above 900 MHz, the system produces major stutters and FPS drops to around 10–20 before falling back again. This creates a repeating, unplayable micro-stutter loop. I only began monitoring clock speeds after the issue started, so I'm not sure whether the 900 MHz behavior existed beforehand.
There was no recent NVIDIA driver update that obviously triggered the problem, and reinstalling Windows did not resolve it. I previously used Arctic MX-4 for about a year without trouble, then applied a Honeywell phase-change thermal material four months ago. It worked well until this issue appeared. Afterward, I repasted the laptop with Arctic MX-4, but the symptoms stayed exactly the same.
What could cause the GPU to throttle this way after a delay, and what should I test next?
2 Answers
This sounds more like a thermal or power-management problem than a simple game setting. The CPU and GPU share the laptop’s cooling system, so heavy CPU boosting can consume the available thermal headroom and force the GPU to throttle. Try temporarily disabling CPU boost or reducing the CPU power limit, then check whether the GPU can maintain higher clocks and power without the stutter loop.
Also inspect the fans, heatsink fins, and airflow for dust or a failing fan. The VRM area may also be overheating even if the reported CPU and GPU temperatures don’t look extreme. A bad heatsink contact or degraded thermal pad can produce delayed throttling, so monitor CPU package power, GPU power, hotspot temperature, fan speed, and the actual throttle reason while the problem occurs.
A driver reinstall is still worth testing if it is done with Display Driver Uninstaller in Windows Safe Mode, followed by a clean installation of the NVIDIA driver. However, since Windows itself was already reinstalled, this is less likely to be the root cause unless the same driver or laptop control software was automatically restored afterward.
Before changing more software, reproduce the issue while logging the GPU temperature, hotspot temperature, power draw, clock, voltage, fan speed, CPU package power, and thermal or power-limit flags. If the GPU is being forced to 900 MHz by a thermal or power limit, the logs should identify which limit is active. If temperatures and limits look normal, test with the suspected modified game executable removed and use a known-good installation, since untrusted software can affect system behavior even if it does not directly damage the hardware.
That’s what confuses me: the laptop was cleaned and the thermal material was replaced four months ago, and the issue appeared suddenly. I also tried a newer game executable with a DRM bypass shortly before this started, so I’m wondering whether software could have triggered it.