|
January 26, 2017, 12:25:28 PM |
|
OK, I have bought a new batch of GPU's there are all RX470 Nitro 4GB with samsung memory. Total 6 GPUs in one Rig.
No matter what I do always 3 cards freezes after 0 up to 2 hours.
Here is what I have done:
- Exchanged PSU, MB, RAM, Risers, Processor.
This lead me to conclusion this is not hardware related problem so I messed with GPU bioses, figured I cut to much voltage so came back to no offset, changed timings (all kind of modification, 1500 , then 1625, then 1750 ended with 1900). Using 1080 core and memory from 1950 up to 2100. Everything same. So came back to stock bioses, here is a big surprice: nothing changed, always GPUs freezes after a while.
Then I splited GPUs to 2 separate machines, whenever there are up to 2 GPU's in a machine everything looks to go smoothy, whenever add 3rd GPU one of them starts to crash.
Oh one more thing, whenever I got connected 6 GPUs, and lets say number 4 freezes, as soon as I unplugged it the problem looks to be shifting to next GPU, and it is not because of different numeration very often it goes to 0 or 1 for instance (which run smoothy with previous configuration).
Also thought it might be a problem of GPU itself, however whenever I connected the worse one to Rig where is up to 2 GPUs connected they run for hours without a single crash.
I don't have any other ideas and don't know what could be a problem or what should I check. Any help would be appreciated.
|