Sli one card is hotter then other

ive got the CLEVO P870DM-G and the sli set up is two gtx 980ms 8gb

but everytime im playing a game or running a benchmark one of the gpus is always running hotter then the other one, 1st gpu is runs at 60'c to 65'c and then 2nd gpu runs at 81'c to 87'c which has me worried....what should i do ?

 

Sounds like the hot one might need a repaste

 

ive repaste them last month

 

Swap them round, does it follow or does it stay in position?

 

thanks for the reply mate...what do you mean by that

 

right now I've got it running heaven benchmark for like 20 mins so far and GPU 1. 60'C and GPU 2. 87'C keeps dropping to 86'C and 85'C....no shutdowns or crashes so far

but in some areas of the benchmark im getting both GPUs running at 987mhz
and their memory usage is at 1477mb on both

 

Was this temp difference present before the repaste? Did it start right after the repaste? What did you repaste with?

Check the alignment. I found it incredibly easy to bend the fragile long heatpipes when I was working on my P870DM.

 

it was the same mate thats the whole reason why I repasted them, yeah id have to check that out

 

Pull out the heatsinks and take some pics of the thermal paste distribution on the GPU die + on the heatsinks and post them here. We'll be able to see if there's any large gap between the heatsink and the GPU die. Also, once you cranked the fans to full, can you check the back of the laptop to see if there's hot air coming out? Could be that the hot GPU had a clogged vent, or that the its fan is failing (less likely).

 

will do that now mate

 

I have done what you told me to do and uploaded the pictures and also checked out for fans blowing out heat which they do. I have repast the CPU and the GPUs but same results to be honest the cpu and the one gpu have lowered few degrees down but nothing too major but still good but the other gpu still running hot up to 87'C

 

Your (slave?) GPU has pretty bad thermal paste coverage. The top 1/3 of the die has almost no paste on it.

 

I have repasted it again but the issue still there mate

 

That looks like a heatplate or thermal pad issue preventing good core contact, then.

 

thanks I appreciate the replies and how can I fix this issue then?

 

A couple things you can check:

1. Apply your normal amount of thermal paste and screw on the slave GPU heatsink with the thermal pads removed. If it makes good contact, then thinner pads in the problematic area should fix it.
2. If the issue persists with the pads removed, try checking the flatness of the heatplate using a ruler.

 

The paste looks disturbingly uneven. Did you reseat the thermal pads after the first repaste before all the overheating issues with the slave card? Also, can you check the mounting screws on the slave card to see if they are worn or something? Bad thermal grease spread could be from not enough pressure on the 2 screws on the lower side.

 

The pads are wrong, there is no contact on the VRMs.

 

Better lay the thermalpads on the card to make sure they are all in right place. I've noticed that the long blue strip for mosfets are on top of the other thermal pad making a thick spot. This probably is the cause.

 

You can bend the four springs to increase the mounting pressure. Also when you screw it down apply pressure to the heatsink to make sure it is making better contact.

 

Check the alignment with a dry fit (take all the pads off, take all the paste off)
It should contact squarely on the square retainer (?) around the core without significant contact between the heatsink and the chassis anywhere else. There should be no rocking or wobbling back and forth when you press in various places on the heatsink.

Blow those dust bunnies out of the rads while you're there

Put the pads back on by sitting them on the card and placing the heatsink down. The heatsink should grip the green sticky pads for good contact. The not-sticky blue ones are harder to judge, the presence of indentations on the underside shows good fit

The VRM pad on the slave also looks a bit wobbly. It has to be straight or it might have been partly caught between the heatsink and inductors and that would cause the back end of the card to sit too high ... which would cause the VRM/memory side of the core to sit high and have less (or no) contact which is what your pic shows.

 

and wtf is going on with your VRM thermal pads on the master? there looks like no thick strip for any VRM contact whatsoever?

 

Thanks for the replies guys I’m here thinking to buy new thermal pads but don’t know where to buy them instantly from the uk I really don’t want to order them online because I need to resolve this issue urgently possibly today

 

Is there any shop where I can get them and also for this gpu what size and thickness of thermal pads do I need

 

Overclockers do thermal grizzly pads in various sizes that you can stack.

 

thanks what size for the gtx 980m or should i still use the ones i got installed?

 

3mm, 1mm and 0.5mm is the mix iirc.

 

more depth please

 

https://www.google.com/search?q=p870dm+thermal+pad+layout

 

ive done what i could but still problem is present.. but i checked other people videos on youtube with gtx 980m sli and their second gpu does go up to 80c

 

What does your contact pattern look like now?