2 GPUs out of 6 in the rig works at abysmally high temps?

[clt]

Hi to everyone. I got a 6 ASUS STRIX GTX 1070 rig all connected with powered risers. When I mine, temperatures of 4 GPUS starts at 45 C and stabilizes at like 60 C, but the other 2 have this steep temp curve starting at 45 Celsius and rise upto 82 C. These 2 gpus also do not take any form of overclock. They can't even mine at %100 TDP with no overclock and keep throwing cudamemcpy error. They only can work at %75 TDP with no oc without throwing errors and sometimes they hash abysmally low like 200 Sol's for Equihash. I changed the risers, cables and did the test and control group thing but had no luck. SO what could this be about? Probable dead Ram, GPU or an issue with PWMs?
Any help is appreciated.

[leonix007]

you can do series of test to help you find the culprit

start with installing crampy GPU's exclude the other 4

Don't use USB risers, connect the GPU's only with PCIex16

if you have 1x PCIe slot, then install only 1 GPU

fireup the miner and observe

if problem still persist, I guess you have the faulty card

[edwardceng]

it's the same problem with my AMD card and I can be fixed some way like this:
■ Reinstall mining software
■ Replace the processor paste on the GPU (this is the final solution)

but is it the same logic to apply in NVIDIA to solve problems like I do on AMD cards