Measuring GPU temps: who is more accurate?

They make noise, too.

Moderators: NeilBlanchard, Ralf Hutter, sthayashi, Lawrence Lee

Post Reply
FartingBob
Patron of SPCR
Posts: 744
Joined: Tue Mar 04, 2008 4:05 am
Location: London
Contact:

Measuring GPU temps: who is more accurate?

Post by FartingBob » Sun Apr 20, 2008 10:30 am

I currently run my 7900GT passively using the Accelero S1 (with AS5 gloop) in an antec SOLO.

Because im running it passively, i like to keep an eye on the temps, make sure they dont get too toasty. My complete build and fan setup is all here so i wont post it all again.
Anyway, Below are GPU temps as reported by
  • nTune 5.05
  • ATItool 0.26
  • Everest 4.0
Ambient temp 28c

IDLE 1 - Exhuast fan (1200rpm slipstream) on full 12v
nTune: 43c
ATItool: 37c
Everest: 44c

IDLE 2 - Exhuast fan at 5v
nTune: 46c
ATItool: 41
Everest: 46c

LOAD - Using Geoforms, max settings for 20 minutes, fan at 12v
nTune: 56c
ATItool: 49c
Everest: 57c

LOAD - 5v
nTune: 67c
ATItool: 61c
Everest: 67c

Why does ATItool consistantly show a 5-7c difference compared to the other 2, and which is more accurate?

I know SPCR use ATItool for their measurements, as do other sites, but ive seen nTune used on numerous occasions as well, and since its made by nVidia i would expect it to be accurate on a nVidia GPU.

My temps are good enough that i dont have to worry about overheating, indeed i took off the 500rpm slipstream i put on the S1 when i installed it because it took up too much space and found temps were still well within safety limits. The fact that i have 2 92mm intake fans facing it certainly helps, the bottom one in particular lines up nicely with the S1.
But still, it would be nice to know im getting the correct reading.

andyb
Patron of SPCR
Posts: 3307
Joined: Wed Dec 15, 2004 12:00 pm
Location: Essex, England

Post by andyb » Sun Apr 20, 2008 11:20 am

My reasoning is very simple, go with the highest one add a few degrees and you are well within limits.

Either way you have ~30C worth of overhead if everest is to be believed with the system fan at 5v. Dont worry, keep on Farting, Bob.


Andy

bonestonne
Posts: 1839
Joined: Mon Feb 05, 2007 2:10 pm
Location: Northern New Jersey
Contact:

Post by bonestonne » Sun Apr 20, 2008 12:14 pm

:roll:

you might try GPU-Z from techpowerup.com

GPU-Z and ATi-Tool give me the same readings on a 9600GT.

yacoub
Posts: 139
Joined: Sat Jun 11, 2005 7:49 pm

Post by yacoub » Sun Apr 20, 2008 12:48 pm

GPU-Z reports a temp several degrees lower than the NVThermalDiode. I would also recommend taking the safest course, which is to go with the highest reported temp.

FartingBob
Patron of SPCR
Posts: 744
Joined: Tue Mar 04, 2008 4:05 am
Location: London
Contact:

Post by FartingBob » Sun Apr 20, 2008 2:57 pm

Tried out GPU-Z and it indeed shows the same temps as ATItool.
And as i said in the original post, my temps don't bother me, but it would be good to know which monitoring program gives the correct value, and why some differ when presumably they all read from the same sensors.

Plekto
Posts: 398
Joined: Tue Feb 19, 2008 2:08 pm
Location: Los Angeles

Post by Plekto » Sun Apr 20, 2008 9:55 pm

Everest looks to be the best of those. It reads the internal CPU sensor that controls throttling and applies a formula based upon how close to that point it currently is.

A load of 57C is typical, actually. 67C is too steep. I'd run the fans at 9V at see if that doesn't give you an acceptable 60C or so.

batka
Posts: 75
Joined: Thu Aug 16, 2007 6:53 am
Location: Budapest, Hungary
Contact:

Post by batka » Fri Apr 24, 2009 5:34 pm

Hi!

I've noticed that GPU-Z shows 3 temps on my Radeon HD 4670:
GPU Temperature
GPU Temp #1
GPU Temp #2

Which temps are meaning which?

Thanks!

Post Reply