What caused the stability issue with my GPU OC?
What caused the stability issue with my GPU OC?
I own an EVGA 1080Ti FTW3 with a stable overclock at +120 on the Core clock. I maintained that configuration for several months without any issues. However, one day my performance dropped significantly during gameplay and benchmarking, and I couldn't identify the cause since none of my settings changed. Testing with 3D Mark revealed that disabling the Core clock OC resolved the problem, indicating the memory OC wasn't the issue. Additionally, reducing the core clock multiplier didn't improve stability—it only extended the time before crashes. If anyone has insights or solutions, I would really appreciate it.
There are several parts inside a graphics card, similar to those on your motherboard except for one key difference. Unlike the motherboard which uses multiple temperature sensors, the GPU only has one. There are no temperature sensors on the VRM, VRAM, PCIe slots, or USB connections. This makes it straightforward to notice that the VRMs can reach 95-100°C, while the GPU only shows 70-80°C. These hidden temperatures can damage the card if not addressed. A minor cooling issue on the heatsink near the back edge of the GPU can quickly turn stable voltages into overheating problems.
Sadly, despite a thorough uninstall and reinstall, the problem persists, indicating it's not related to drivers. Appreciate your effort in helping.
Capacitors on your graphics card and power supply unit may react poorly to increased heat, reducing their ability to filter ripple effectively and potentially causing an overclock that was previously stable at lower temperatures. You might need to boost fan speeds, adjust voltages, enhance air conditioning, or lower the overclock.
Thank you for your efforts, but the stability check passed and temperatures stayed consistent. I was already operating at full voltage, the fan curve didn’t behave as expected (my GPU reached around 50°C under load, but pushing fans to max didn’t significantly lower it—just a few degrees), my room is about 21°C, so I don’t need to adjust the AC just for this. Also, from your original post, you noted that reducing overclocking isn’t beneficial. There isn’t a single stable overclock, even small ones like +25 don’t work.
The PSU that is overheating is more likely to be the source of the issue. What model is it?
Even so, aging capacitors are unavoidable, but they can be reduced by maintaining a cool environment. The dielectric breaks down over time, which explains why PSUs with heat-resistant caps tend to have longer warranties.
It's surprising how many PSUs exist only rated for 25C-30C by major brands such as EVGA and Corsair. These caps tend to burn out quickly, often failing to meet the 2-year warranty. Poor-quality caps are also common on budget motherboards and graphics cards, but they pose a greater risk with PSUs.
The filter circuit plays a crucial role in maintaining overclock stability. You should aim for a clean DC output—avoid anything like this ~~~~~.
There are several parts on a graphics card, similar to those on your motherboard except for one key difference. Unlike the motherboard which uses multiple temperature sensors, the GPU only has one. The graphics processor unit (GPU) lacks temperature sensors on its VRM, VRAM, PCIe, or USB components. This makes it straightforward to notice VRMs reaching 95-100°C, while the GPU only shows 70-80°C. These hidden temperatures can damage the card if not addressed.
A minor cooling issue on the heatsink near the back edge of the GPU can cause stable voltages to turn into overheating conditions. It might be helpful to use a small Phillips screwdriver, remove the cover, and thoroughly clean the entire heatsink/fan area. Check that all fans are operating properly.