After reading a lot about 5700 XTs overheating because of cooler mounting, I am concerned for mine. Its an XFX Thicc III Ultra v2. When I play titles like Frostpunk, the GPU temp sits in the mid 70s, but the ‘hot spot’ temp always hits 100. I know this card’s BIOS attempts to aggressively overclock and maybe the temp is its throttle point, but I would like to do some testing and then re-paste and tighten the cooler to see if there is any difference.
That said, what is the Prime95 of the GPU? What software should I use for thermal testing?
Furmark… It’s a power virus. It literally tortures your GPU to use as much power as possible and therefore dumps as much heat as possible…
Oh damn, they have V2 now…
The original 5700XT was Thicc II. After all the complaints they made Thicc 3 with improved everything and now there is a revision of the Thicc 3.
That’s not so bad…
Please tell me that’s the actual GPU and not memory or VRM if they even have VRM thermal sensor.
It’s sad how unexcited I am about XFX at the moment. Their RX400 cards was suuuuuper amazing. Their R9 200 line was also good, but alas, we reached the point where XFX are less exciting than freaking PowerColor…
Anyways, what is your case? Your ambient room temperature?
Yes, its the GPU junction temperature that sits at 100. The VRM doesn’t break 70 and the Memory sits around 80.
The case is a Lian Li PC-A71F from 2009. The room temp is a balmy 19C (yay winter).
What I have found is the XFX fan curve sucks. If I leave the fan under full automatic control, it won’t exceed 60% and instead the card throttles as the junction reaches 100C and the memory hits 90. I turned on the manual fan control and allow it use the full 100% of the fan curve and everything changes. Now as the junction temp goes up so does the fan. When the fan reaches 100% the junction temp now stays below 90 (85-90) while the other GPU sensor stays in the high 60s. In this mode the VRMs stay in the 60s and the memory about 80.