Your frame rate has decreased noticeably.
Your frame rate has decreased noticeably.
Great, thank you for your feedback. I’ll focus on what you need: capturing the initial system state after a power-off, showing startup metrics, game boot behavior, and post-game shutdown visuals. I’ll ensure the screenshots include the necessary charts and that the timing aligns with your instructions.
I made an error believing you were interested only in the final graphs; here are the additional ones. These are the initial visuals for when the game is running and these appear once the session ends.
Great work overall. The information provided is quite helpful. The emphasis on GPU power consumption is solid. In summary, the issue seems to be related to the GPU's cooling system. It’s experiencing significant thermal throttling. Noticeable drops in clock speeds occur, and only a fraction of its power capacity is utilized—unless a FPS cap is active (like with VSync). The memory temperatures appear normal, which suggests potential problems with thermal pads being too thick or the GPU module not being installed correctly. Or perhaps there’s a compatibility issue with your specific card. I can elaborate on my reasoning if you’d like.
I understand how you reached your conclusions. Since VSync isn't active and everything was already set up for the waterblock, I wonder if reapplying the thermal pads is necessary. I can reinstall the waterblock to confirm compatibility with my GPU, or perhaps a different radiator could help with the thermal throttling?
If the block works and includes the right thermal pads, the main problem likely lies with the GPU cold plate attachment. If it’s installed properly, idle temperatures for both GPU and CPU should stay close together—about 1-2 degrees apart, and roughly 0-5 degrees above the coolant temperature. After fixing that, consider adding another radiator to the loop, but first ensure the GPU cold plate is secure. Remove the GPU, observe the thermal paste texture (photograph if needed). Regarding your Windows setup, running in high power mode with Nvidia Control Panel’s “Power Management Mode” set to high performance can be efficient. It’s better to adjust the NCP settings per game instead of using a global high-performance setting, so the GPU can downclock when you’re idle. For the Windows Power plan, balanced mode is usually a good choice—it lets the CPU downclock while saving power and reducing heat compared to always-on GPU operation.
I was thinking about dismantling the block today, but when I checked the GPU and CPU idle temperatures, they matched what you expected with the cold plate installed properly. The GPU is idling between 38-40 and the CPU around 40-42. It’s still throttling slightly but not as severely as before. You also adjusted the power settings as recommended. Should I still go ahead with the dismantling, or could there be another issue? Let me know if you’d like me to share the graphs again.
The hot spot temperature was also higher than expected for the GPU. Usually, the difference should be about 10°C, but yours was near 30°C or more. This suggests uneven cooling of the GPU die, pointing to a possible mount issue. Feel free to share additional graphs if you'd like. I'll review them and let you know my thoughts.