My Radeon 6800 XT overheats and loses connection to the computer while gaming, often crashing with a GPU timeout even in relatively undemanding games. During a stress test, the hotspot reaches 110°C in about five seconds. I've reinstalled the drivers, reseated the card, and checked the case airflow, but nothing changed. The only reliable workaround so far has been underclocking, undervolting, and using a more aggressive fan curve. Could the thermal paste or thermal pads be deteriorating? Is it safer to keep using the card with conservative settings, or should I pay to have it serviced before the problem gets worse? I can't afford to replace the card and would prefer not to disassemble it unless the repair is reasonably safe.
2 Answers
Repasting the GPU die is a sensible next step, along with checking that the heatsink is making proper contact. A phase-change material such as PCM can also work well for the die if you choose a suitable product. Keep using the undervolt and lower clock speeds in the meantime, but if it still reaches 110°C almost immediately after servicing, the cooler, fans, or the card itself may have a hardware fault.
The thermal interface material may have degraded, especially if the card has seen a lot of use. Replacing the paste is probably the most practical option before considering new hardware. It’s a fairly manageable repair if you follow a teardown guide for the exact card model, work carefully, and use the correct amount of paste. If you replace the thermal pads too, make sure you verify their thickness first, since the wrong size can make cooling worse.
Thanks! I was worried that repasting a GPU could easily damage it, but I’ll look up a detailed guide for this exact model and consider doing it carefully rather than paying hundreds for service.