How temperatures affect power consumption

[Cyclops]

I made a video about this subject yesterday. Someone out there might find it interesting.

[Necrodead]

I expected the results but always good to know for sure! Good vid Cyclops.

[gdesmo]

Good video sir !

[DazMode]

So good cooling pays for itself after 1 year with ontario hydro prices!

[Cyclops]

Haha yes... tell that to air cooling guys with temps in the 80s!!

[Bartacus]

So good cooling pays for itself after 1 year with ontario hydro prices!

Counter balance that with the good fans you need for rads. My GPUs never hit 40C, but there are 17 good fans on my rads. I'm guessing it all balances out in the end. We do get arse raped for power in Ontario though!

[MemBusX]

Very interesting Cyclops. I've never heard of this until now. I should try testing that with my 780 on air. I've got a electricity usage monitor too from when I was buying a UPS. I wish the UPS software was more reliable for those readings

[MemBusX]

OOOOOOkkkaayy. I got off my ass and did the test to collaborate. Like I said I was gonna do. Pics up soon

[Hooded]

Good tech.

[MemBusX]

I did some testing of my own here. In the name of science and boredom. I said I would try this about 5 months ago, so here it is. Pretty much confirms Cyclops' observation.

Tried as much as possible to be consistent with the Video. Furmark is run for 45 mins. Temps and power consumption is recorded before the computer is turned on, While the computer is idle/cold and finally after the 45 min Furmark test. For the readings I'm using the RioRand Power Meter and just for kicks, I took the readings from the UPS display to see how accurate it is.

Lets see how much power this old and inefficient system uses.

Specs
CPU: AMD FX 9370 (Stock)
GPU: Zotac GTX 780
RAM: 16 GB G Skill Ripjaws 1600 Mhz
MB: Asus M5A99FX Pro R 2.0
PSU: Corsair RM 750 (Gold Efficiency)

This is my Command Center. Everything here is plugged into the UPS then to power meter. Except the desk fan.


This image shows the draw while computer was off. The only thing drawing power was the Router and perhaps the monitors on standby.


Here we have the computer sitting idle right after login. The numbers are different but they were fluctuating very rapidly as the CPU did back ground tasks. They were pretty much ballpark of each other though.


42 mins of power on idle, if power goes out.


These are the readings before the test.


1:25 mins into test



The Readings on Meter at start of test



about 45 mins later. This is what the sensors read


Here's the heated power draw of the System



So yeah. About 30-35W +/- extra draw for a single card system. That's quite a bit.

[Cyclops]

Science

[MemBusX]