[SOLVED] Minimize GPU temp fluctuation by raising idle temperature

[YamatoDamashii]

I've been playing with the fan curve for my rx 580 recently, and came to the conclusion that having a higher idle temperature (45-50C) and lower load temp is preferable to trying to keep it cool as possible at all times.

I play high-intensity games on and off in addition to sporadic image/video rendering sessions. Rather than having my GPU jump from sub 30C temperatures to the high 70s quite rapidly repeatedly, I wonder if it would be better to have a higher idling temperature so that the transition, and hence the experienced stress on the card, is minimized when switching from web-browsing to intensive tasks.

I have the fans turned off until the default kick-in temperature threshold (I think 50 something) is reached. I also aim to maximize fan life this way and minimize dust contamination.

Tell me if this analogy doesn't apply, but I think of a car's engine wanting to be warmed up before you drive it hard. You don't want to cool the engine down so much to where the jump in temperature is drastic between idle and intensity.


Has anyone else here heard of this before? Or am I completely wrong?

 

[Darkbreeze]

Keeping it as cool as possible at all times, whether with a CPU or GPU, is a waste of time. The ONLY thing that matters, AT ALL, is that the maximum temperature does not exceed the recommended specification. For most CPUs that's 80°C and for most graphics cards it is 80-85°C. If you below those numbers, it does not matter at all if you are below them during PEAK, MAXIMUM load, by one degree or by twenty degrees. It will have no effect on performance or longevity UNLESS you start to exceed those peak temperatures.

The GPU core or any CPU core, are fully "warmed up" to use your analogy, by the time you resume from sleep or boot into the windows desktop, as much as they are every going to or need to, be.