Nvidia's GTX 1080 Goes on Sale, Promptly Sells Out Everywhere

[Strazdas]

I don't know what PCper did to keep temps down, but many review sites are easily hitting higher temps. JayzTwoCents seemed to hit 84 degress very quickly when playing Crysis 3. Likewise an excerpt from the HardOCP review states:

Naturally the Radeon R9 Fury X is running the coolest because it is using a closed-loop liquid cooling system. Compared to the air cooled GeForce GTX 980 Ti though the maximum GPU temperature seems to have a similar thermal limit set on the GTX 1080 as it is on the GTX 980 Ti at 83-84c. When it hits those temps it will throttle voltage and clock speed and TDP to keep it in that range. The only way to alleviate this is to use a third party overclocking program to raise the temperature cap. Or you could raise the fan profile to keep it cooler.

yeah but what are the clocks at that 84C thermal limit? because if they are same as advertised core clock then the card isnt actually throttling itself and more the GPUBoost is not turning on at hot cards, which is expected behaviour.

Altrough the limit does seem a bit low, given that some older cards easily could hit 90+ without a problem (titan even reaching 100 sometimes).

Jayz seems to be having a meltdown of sorts nowadays.

 

[TotalerKrieger]

I don't know what PCper did to keep temps down, but many review sites are easily hitting higher temps. JayzTwoCents seemed to hit 84 degress very quickly when playing Crysis 3. Likewise an excerpt from the HardOCP review states:

Naturally the Radeon R9 Fury X is running the coolest because it is using a closed-loop liquid cooling system. Compared to the air cooled GeForce GTX 980 Ti though the maximum GPU temperature seems to have a similar thermal limit set on the GTX 1080 as it is on the GTX 980 Ti at 83-84c. When it hits those temps it will throttle voltage and clock speed and TDP to keep it in that range. The only way to alleviate this is to use a third party overclocking program to raise the temperature cap. Or you could raise the fan profile to keep it cooler.

yeah but what are the clocks at that 84C thermal limit? because if they are same as advertised core clock then the card isnt actually throttling itself and more the GPUBoost is not turning on at hot cards, which is expected behaviour.

Altrough the limit does seem a bit low, given that some older cards easily could hit 90+ without a problem (titan even reaching 100 sometimes).

Jayz seems to be having a meltdown of sorts nowadays.

Sorry made an edit after you replied. From Tom's Hardware:

The 1080 hits its temperature target by dropping the GPU's clock rate. During a gaming loop, it falls all the way down to its base frequency, leaving nothing left of GPU Boost. This gets even worse during our stress test, where the core clock dips below the 1607MHz that is supposed to be the GeForce GTX 1080's floor.

http://www.tomshardware.com/reviews/nvidia-geforce-gtx-1080-pascal,4572-11.html

In a real world gaming situation, clocks can fall from just under 1900 MHz to just above 1600 MHz after only a few minutes.

 

[Strazdas]

Sorry made an edit after you replied. From Tom's Hardware:

The 1080 hits its temperature target by dropping the GPU's clock rate. During a gaming loop, it falls all the way down to its base frequency, leaving nothing left of GPU Boost. This gets even worse during our stress test, where the core clock dips below the 1607MHz that is supposed to be the GeForce GTX 1080's floor.

http://www.tomshardware.com/reviews/nvidia-geforce-gtx-1080-pascal,4572-11.html

In a real world gaming situation, clocks can fall from just under 1900 MHz to just above 1600 MHz after only a few minutes.

The furmark one is the only one that looks worrysome, and thats a synthetic test.

Running in base clock during full load loop is expected bahviour. The boost is meant to be a temporary measure to begin with, even if it sometimes can work permanentely.

The card, running in base clock, does not exceed throttling temperatures. Working exactly as advertised. I dont see whats the big fuss is about. Unless its all about that furmark test (which indeed is a bit worrysome given that the car should maintain stable speed at that temperature even on max load rather than just real world scenario). I guess this is where our reviews differ, as PCPer used Heaven 4.0 as a long term temperature/clock test which is closer to real world than to synthetic test and Tom ram it through synthetics as well.

On one hand this does sound like its bad OC potential though. However as you mentioned there are ways to disable that thermal limit (which is pretty low for a GPU to begin with, but new architecture may require it, who knows) and given that AIBs are going to have better cooling on it i think the OC is going to be allright.