The thermal throttle limit (default) of that card is 83 C I think. But it should down clock though not crash. If you get an app called GPU-Z there is a Sensor tab, in there is a line called PerfCap Reason. This can display some confusing information, but it can help diagnose some thermal or power issues by indicating WHY the GPU is at a given clock speed (this includes throttling and full speed, which is where it gets confusing). This can be Idle, say, when your on the desktop or any low GPU load task, or VRel which is what you would typically see during gaming and is completely normal. If you saw anything related to temperature then you might be on to something. It's a free download pretty much anywhere, I think Asus may even provide a version. Like any bit of third party software make sure it's from a reputable site, and of course scan it before installing etc etc. Just know that any time we look at graphs such as those displayed in Afterburner, Windows Resource Monitor etc we are not seeing the whole story. The refresh rate on these apps can only be so high before they start to consume a noticeable amount of CPU resources. This means any spikes in clock speed, voltage and yes, even temperature often get missed. That's why GPU-Z and similar software which monitor the PerfCap bit are so useful. Didn't mean to bust in the conversation but hopefully some of my nonsense can help?