GPU fault detected

[akinava]

Hi,

I'm starting my very first rig with two RX5** cards, on Linux with Claymore miner.
GPUs start to mine, but after a while Sapphire RX 560 disconnects with following error

==> /var/log/syslog <==
Jul 11 01:52:00 uminer001 kernel: [  821.137517] amdgpu 0000:01:00.0: GPU fault detected: 147 0x0d284808
Jul 11 01:52:00 uminer001 kernel: [  821.137526] amdgpu 0000:01:00.0:   VM_CONTEXT1_PROTECTION_FAULT_ADDR   0x001493A5
Jul 11 01:52:00 uminer001 kernel: [  821.137530] amdgpu 0000:01:00.0:   VM_CONTEXT1_PROTECTION_FAULT_STATUS 0x0E048008
Jul 11 01:52:00 uminer001 kernel: [  821.137536] amdgpu 0000:01:00.0: VM fault (0x08, vmid 7) at page 1348517, read from 'TC0' (0x54433000) (72)

Even if I keep only this GPU connected, it works a little bit longer but still fails.
The card is running with stock settings 

Do you guys have any ideas that could be wrong with it?

[igotek]

the same problem here.
How to fix it on HiveOS ?

[rdluffy]

It's probably a problem related to Linux + Drivers
Try to update your linux and amd drivers (if you're on latest drivers, try one before)

If not works, I suggest you to try Windows or another distro for mining like https://simplemining.net/

[rikuu]

That error is a segfault, meaning that the GPU tried to access an invalid memory address.
Try with stocks settings (no OC profile, no modified memory straps, etc) and see if it still happens.