Temps on gpu's in crossfire

AgedIdiot

AgedIdiot

AgedIdiot

Soldato
Joined
31 Oct 2007
Posts
2,606
Location
Gods County(Allegedly)
Am running 2 x 7950's in crossfire. Am running one loop

Pump(D5)/Res>360(roof)rad> Cpu> Gpu 1 > Gpu 2> 120 rad (floor) >Pump/Res.

Certainly a lot quieter and cooler than under air :)

Somehow managed to lose the Riva monitoring service so set about uninstall + reinstall to get it working again. It works again :)

To test it I ran Heaven and the reason for posting here was to refer to the max temps on each gpu when stressed through Heaven. In summary

Gpu 1 - Max Temp 39c - Vrm's 42c

Gpu 2 - max temp 45c - Vrm's 44c

Given the fact that they are both cooled wth the same blocks I anticipated that gpu 2 might be a little hotter but was surprised that it was as much as 6c.

Normal ?
 
Are they both loaded the same?

Location of blocks makes very little impact on temps, the water will reach temperature equilibrium and location in loop means naff all really.

If anything it will be because the block is not as well seated on GPU2 as it is on GPU1. Or GPU2 is loading more/pushing more voltage.
 
It's not particularly unusual and if you've just done your loop, chances are there's still little pockets of air in the blocks.

My 7970's were a couple of degrees apart most of the time fwiw. A couple more isn't a biggie :)

What clocks/voltage btw?
 
Been running for about 2/3months so would have thought would be settled.

Am running both at stock speeds (900 core + 1250 memory)and volts - left it to Asus (gpu1) and Gigabyte (gpu2).

Will have a look to see what volts are and report.
 
Begining to believe that it's more of a reporting problem with Gpu-z than a 'real' problem.

Amended Riva to report voltage when running - ran heaven again and noticed that the voltage to both was identical 1.094v and the temps were increasing with core 2 adding 1c which was added to core 1 a few seconds later - I believe that if I had run it for long enough they would have settled at same temp.

GPU-z reported that :-

max temps difference was , now, 1c.:)
Volts max was 1.086 for core 1 with a little more 1.088 for core 2 (might explain 1c difference?)
VRM's - a bit of a role reversal as gpu 2 is showing 1 c less on vrm1 AND 3c less on vrm 2.

Overall , unless someone advises me to the contrary I will assume that any differences are either acceptable or miss-reporting :) Makes sense?
 
my temps with 2500k at 4.6 and both my 290x at 1100/1375 with 2 rads one being a 3 fan and 2nd 2 fan my temps are 70c on gpu1 and 65c on gpu2 after running bf4 for a hour. all fans on max speed of 1500rpm
 
I just transferred a couple of 290's into my main rig today (360+240 rads) and they're only in the 40's, maybe pushing 50 depending on how I choose to torture them :p

I think you may want to have a look at your setup, Jordan :) Might just need a good dust/vacuum if my past laziness and resulting temps are anything to go by ;)
 
I will double check but I think my 290X's (stock) hit 52 after torturing them for 30mins or so. I have 2x480's 1x360 and a 250 though
 
You must log in or register to reply here.
Back
Top Bottom