Clock speed reduction causes performance issues; latest GPU model didn't resolve the problem
Clock speed reduction causes performance issues; latest GPU model didn't resolve the problem
I've experienced a problem where my GPU core clock utilization drops sharply for brief periods, causing frame time spikes—sometimes exceeding 100ms. This occurs across various games like TF2 and RDR2. Temperatures remain normal, CPU usage stays stable, and RAM usage isn't excessive. I've tried several fixes: adjusting MPOGPUFIX settings, disabling AMD ULPs in both MPOGPU and registry, reinstalling DDU multiple times, tweaking CPPC in BIOS, modifying GPU power-saving options, setting a minimum core clock in AMD settings/MSI Afterburner, changing pagefile size to a fixed value, replacing the GPU, adjusting voltage profiles, overclocking or undervolting, updating BIOS and chipset drivers, switching audio devices/drivers, using latency monitoring tools, and running OCCT with CPU, GPU, and memory together. Core clock stayed consistent during the issue.
Potential causes might include RAM or power supply problems (though voltage readings looked normal in OCCT/hwinfo). If this happens on other systems, it could point to driver or configuration issues within Windows rather than hardware faults. I'm open to benchmarks or detailed logs—please let me know if you'd like to proceed with that.
Specs: OS – Windows 11 64-bit; PSU – ASUS ROG 1000W (two cables); MBO – MSI B650-P; CPU – 7800X3D; GPU – 7900X; RAM – DDR5 6000MHz CL30, 2x16GB; Storage – 2 NVMe + 2 SATA; Cooler – Deepcool AK620.
Benchmarks:
- http://www.3dmark.com/fs/32970148
- Updated note: Doesn't occur in Linux (PopOS), suggesting a Windows-specific driver or setting issue.
- Updated note: Valve Index USB hub caused the problem; disabling it resolved it.
It's unclear what the benefit would be, but you might consider leaving the GPU settings lower than maximum to provide some headroom.
Adjusted performance to 80% of the standard rate, but core clock interrupts remain. Appreciate the advice!
The interruptions appear consistently across most games I play, including offline sessions in TF2 where no network or bots are active. It seems unrelated to CPU, disk activity, or temperature readings—they seem unpredictable. I tested this in Red Dead and still experienced significant frame rate drops. It doesn’t look like graphics settings are the cause, since similar issues occur in less demanding titles like TF2, CS2, and Rocket League where my hardware usage stays below 100%. The GPU frequency didn’t drop noticeably, and I didn’t see any performance problems during operation.
I installed PopOS on another drive and the performance issues during gaming on Linux resolved. It seems the problem might be related to a driver or configuration in Windows, possibly causing interrupts that disrupt the GPU. I’m not sure what to do next except consider a full Windows reinstall.
I used a Valve Index and turning it off resolved my problems. Additional information was refined in the original message.