Hello everyone.
I have two PCs with the exact same hardware inside them. (let's call them
PC1 and PC2)
CPU: i5 6600
RAM: FuryX 2133 16GB
MOBO: ASUS B150M PRO Gaming
SSD: Samsung nvme 960
PSU: Corsair TX550M
GPU: EVGA 1060sc 3GB Same monitor (Dell 1440p 60hz), same connection (Display port), same windows 10 pro, same drivers.
PC1 is working perfectly.Temperature on the 1060 is around
35c with 0% fan and ~
24c with 40% (idle @ desktop)
(room temp is always 22-23c max) Now the problem:
The GPU on the
PC2 was overheating like crazy.
55c+ even on desktop with nothing running in the background.
Clocks @ 253mhz/650mv usage 0%In-game temp was more than 85c with 100% fan (Clocks less than 1600mhz). After a while I could see artifacts so I used the rma process and a week later a
brand new 1060sc was in my
PC2. Quess what. The brand new 1060sc on PC2 is idling around
55c with 0% fan and
35c with ~40% again.GPU temp is always more than 35c no matter what. Windows boots in less than 30seconds and the card is around 35c when the exact same time the GPU temp on the PC1 is 22-23c.
Load temp is more than 85c in RE7 Biohazard (thermal throttling) when the other card is always around 65c with full boost clocks @ 2012mhz on the same game/same time.
I've tried to: change motherboard, cpu, format windows, other ssd, other bios on the gpu, new TIM... nothing, same temps.
Then I installed a waterclock (EKWB) and used my custom setup with only the 1060 in the loop.
Water temp 20c
GPU temp idle: 30c
GPU temp HEAVEN bench: ~32c after a full pass.
My conclusion is that
some pascal cores are not measuring temperature correctly or nvidia/evga forces an offset (my card is around +10c delta always) for some reason. (maybe boost related?) I've seen a LOT of people complaining about idle temps, is there any possibility for an official answer about this?
Thank you and sorry about the long post.
tl;dr Pascal temperatures are not accurate on some cards for a reason.
post edited by cmosfxx - 2017/02/23 23:15:58