This should fix the burned cards.

I was hit hard by this issue, you know, my fan speed was not going higher than 20% which is default. Due to this unfortunate thing, my GPU temperature raised up to 47 degrees centigrade under load. Why there is even custom fan control when all one needs is 20% of fan speed and silence. People have reported GPUs dead to heat while running water cooling and reaching whooping 60C. Cowie and some other members did test it and their unfortunate cards did reach high temperatures (around 100C), but did they die? No. Does GPU die in matter of minutes/hours/days upon reaching such temperature? No. There are usually many signs like artifacting and many shutdowns before GPU takes final hit. What can die is VRM, did someone report high VRM temperature? I could not find anyone here. Is there chance for GPU fatal blow to happen in just one gaming session? Extremely small, since VRMs would reach critical temperature before GPU, and Voltage drop would cause GPU to shut down anyway.

Omg, dude when the R9 series was released AMD advised that their cards are thoroughly tested at 94C, which happens to be the limit at which Powertune kicks in and throttles the clocks so that the GPU does not exceed this temp. Those who do exceed it are either lying or have a custom bios installed. Any problems those with a custom bios have are their own fault. Running CFX myself my top card will, in a demanding game, hit this threshold, thus I will often underclock slightly in order to remain below it. But often I game for hours (days even) on end with my top card in the 80's. I've had the top card for nearly 2 years. No probs whatsoever, still under the 3 year warranty too. Worrying about 47C is just being paranoid. 60C is not whopping either. Sorry to sound harsh but just trying to let you know, those temps are not/should not be a concern.

Omg, dude when the R9 series was released AMD advised that their cards are thoroughly tested at 94C, which happens to be the limit at which Powertune kicks in and throttles the clocks so that the GPU does not exceed this temp. Those who do exceed it are either lying or have a custom bios installed. Any problems those with a custom bios have are their own fault. Running CFX myself my top card will, in a demanding game, hit this threshold, thus I will often underclock slightly in order to remain below it. But often I game for hours (days even) on end with my top card in the 80's. I've had the top card for nearly 2 years. No probs whatsoever, still under the 3 year warranty too. Worrying about 47C is just being paranoid. 60C is not whopping either. Sorry to sound harsh but just trying to let you know, those temps are not/should not be a concern.

Due to this unfortunate thing, my GPU temperature raised up to 47 degrees centigrade under load. Why there is even custom fan control when all one needs is 20% of fan speed and silence.

People have reported GPUs dead to heat while running water cooling and reaching whooping 60C.

Cowie and some other members did test it and their unfortunate cards did reach high temperatures (around 100C), but did they die? No. Does GPU die in matter of minutes/hours/days upon reaching such temperature? No.

here are usually many signs like artifacting and many shutdowns before GPU takes final hit. What can die is VRM, did someone report high VRM temperature? I could not find anyone here.

Is there chance for GPU fatal blow to happen in just one gaming session? Extremely small, since VRMs would reach critical temperature before GPU, and Voltage drop would cause GPU to shut down anyway.

47C is barely warm...;) You should make up your mind...either you run your gpu "under load" or you run it on 20% "silent" fan speed. Not possible...watercooled GPUs have no fan and would not be affected. Anyone who would knowingly restrict his fan to 20% and then "test" it under load to see if it gets hot is an idiot...;) There can be no "final hit" because GPUs for years have been made to shut down long before they might burn out due to temperature. Sorry, but no chance of that at all. Hasn't been for many years. GPUs will shut down first, before any damage can take place. Your post does not indicate much experience in these matters.