Weird GPU boost behaviour

ThatSput · April 20, 2016

Hi everyone. I am not actually sure if this issue has anything to do with NVIDIA GPU boost. It is more likely my PSU, but this is so confusing that I need your help.

Basically, my 1,5y/o graphics card (EVGA GTX 780 SC w/ ACX) started to run at much lower clocks, with minimum GPU boost. The thing is that the card is not throttling because of being too hot, the temps are totally fine. What's even weirder is that it keeps switching between "working totally fine" and "now I have to chill again". One time I get max boost with max voltage, then randomly it shits the bed and runs at almost idling voltage, despite 100% GPU usage and temps being far from the limit.

I took some pictures of GPU-Z readouts.

https://imgur.com/kLeW17q

Temps are fine, max load (Heaven benchmark), yet dropped voltage/core clock and low TDP.

https://imgur.com/M6ZnQjX

Temps still fine, this time max boost I used to get, max voltage.

https://imgur.com/dZsaBzD

Yet in Witcher 3 still slight throttle despite temp limit is not hit?

First I thought my PSU was dying, but I see no system instability and can stress-test CPU just fine. Clean install of drivers, MSI afterburner settings on default... I honestly don't know what is wrong here. Thanks for any help!

edit:

specs: i7-4770k, 16gb RAM, EVGA gtx 780 SC, ASUS z97-a, Seasonic 650W

Also, all of the pictures are taken after at least 15 minutes of running Heaven/Witcher 3 without vsync