GPU clock speed fluctuations occur occasionally.
GPU clock speed fluctuations occur occasionally.
I increased the overclock of my RTX 5080 GPU by 500mhz, kept the memory clock steady, and after that my PC would freeze. When I went back to normal speeds, the clock jumped from 2600 to around 300-400 and then back down to 2600 again. All connected devices—mouse, keyboard, headphones, two monitors, microphone—worked fine. The PC specs are as follows: motherboard Asus Rog Strix X870E-E, CPU Ryzen 9950X3D, GPU Zotac RTX 5080 Amp Extreme Infinity, RAM 64GB G.Skill Z5 Neo, SSD 1 x 2TB Samsung 990 Evo Plus (1.6TB free), another SSD 1 x Crucial T705 2TB (still unused), power supply Lian Li Edge 1000 watt 80 Gold (new), cooling system Lian Li Hydroshift-C TL AIO and Lian Li TL fans, and I’m using a wireless network. There was no particular cause. Windows 11.
Update your post with complete hardware details and operating system information. Provide the power supply specifications (PSU: make, model, wattage, age, condition—original, build, refurbished, used), disk drive specifications (make, model, capacity, current usage). List all connected peripherals. Indicate whether the network connection is wired or wireless.
Reflect on your reasoning for overclocking: assess whether it was necessary and evaluate any potential benefits in specific scenarios. Note if there are relevant manuals or references that apply to this setup. Consider the overall limitations observed, as the system seems unsuitable for high-level overclocking.
PS: Additional context would help clarify the situation.
My recommendation is to employ Task Manager, Resource Monitor, and Process Explorer to assess system performance. Use all three tools sequentially but one at a time. Initially, you should monitor without overclocking and continue with regular activities like working or gaming. Keep the tool windows open for continuous observation. Next, increase the overclock settings and observe any changes or issues during spikes. Take your approach slowly and carefully, avoiding simultaneous adjustments. The goal is to identify a situation where "when X – there are spikes, when no X – there are no spikes." Or vice versa: absence of spikes without the presence of X. The key lies in determining what triggers these conditions. Pay attention to error messages, warnings, and any data logged in Reliability History/Monitor or Event Viewer during spike events.