Why does 100% GPU fan not do much?

[Guest Console Peasant]

50% is only a couple of degrees more (during winter its nothing) and is so much quiter, why does 100% even exist? (Let me show you in MSI afterburner)

At the middle of the line is where i put it to 100%

 

http://gyazo.com/833477dfc4bb8b05f1173e6e6cf7215f

[blu4]

To see cooler efficiency? Use it when you BIOS hack the GPU? 24/7 operation?

[ind3x]

are these numbers at low-medium usage?

At peak load it would give a bigger difference

[W-L]

50% is only a couple of degrees more (during winter its nothing) and is so much quiter, why does 100% even exist? (Let me show you in MSI afterburner)

At the middle of the line is where i put it to 100%

 

http://gyazo.com/833477dfc4bb8b05f1173e6e6cf7215f

 

it depends on your cooler and there is always going to be a 100% which is the max capacity your fan can spin.

[MrKickkiller]

Card manufacturers know that people want silent cards, so they code the software in such way that it will try to hover around a specific RPM.

About the "Why is there 100%": The fan motor may be capable of 5000 rpm, but then it isn't pushing much more air trough the heatsink.

To keep the silence, the fan is set on a profile that makes it run at a speed of 0- 3500 rpm. This is also better for the fan motor.

[Guest Console Peasant]

are these numbers at low-medium usage?

At peak load it would give a bigger difference

ill do another test while benching witcher 3

[Guest Console Peasant]

are these numbers at low-medium usage?

At peak load it would give a bigger difference

http://gyazo.com/9b85f5ab0b7b6ddfa1f4283b2a06952b

 

started at half way

[REX.exe]

Maybe cuz you're a console peasant?

[Guest Console Peasant]

Maybe cuz you're a console peasant?

shit

[Blebekblebek]

what are you running during the spin test?

[Guest Console Peasant]

what are you running during the spin test?

what is a spin test?

It's a temp test. I was running witcher 3 max settings

[Blebekblebek]

you are testing 100% speed fan? if that isn't called spin test, then what is that supposed to called?

 

anyway, you probably should do synthetic benchmark and just see change from it (from 50 to100)

because benchmark more consistent.

[Snadzies]

Air coolers are all about transferring heat from the heatsink to the air passing through the fins.

Slower moving air gets a little more time/opportunity to have heat transferred from the fins to the air than faster moving air does, however faster moving air usually means a greater volume of air that the heat can be transferred to.

You eventually start getting diminishing returns where even though you're pushing great volumes of air through the fins the air is moving past so quickly that it has little opportunity to have the heat transferred to it.

 

Another important factor is the airflow of the case, the intake and exhaust.

If you're not getting enough cool air in and hot air out speeding up the fans on the GPU will only help so much as the graphics card with aftermarket coolers will basically be "rebreathing" the same air it just heated up a moment ago and blower style coolers will get starved of air if there is not enough air coming into the case.

[ind3x]

The heat pipes can only get as cold as the air going through the fins. 

Let's set this temperature to 20 degrees celsius.

The heat pipes run along the fins, and end at the die. The top of the die is touching the heat pipe which is 20 degrees celsius.

But inside the die it's 50 degrees celsius, but it can't transfer the heat quicker, and therfore stays at 50. 

The temp monitor isn't sitting on top the die, it's inside it. 

I think this is why you aren't seeing any temp differences. The cooler is already peaking in performance at 50% fan speed. If you were to raise the ambient temperature to let's say 30 degrees, then it would have to use maybe 65% fan speed to keep the same temperature. That's why there's room for more speed. It's simply hotter in other areas of the world.