GPU are stopped. Attemping Restart

[smokinggun46]

Rig: Ubuntu 17 64-bit, 4 Gigabyte Windforce OC 1070s , all molex risers, 2 risers powered by molex to 6 pin, other 2 powered by molex to sata then sata to 6pin.

Within 30-75 minutes of mining, EWBF Cuda Miner starts giving the error (paraphrasing): GPUX are stopped. Attemping Restart → GPUX could not be restarted. Please reboot. Sometimes a GPU might be restarted and continues to mine. This morning, the first time I left it on for 7+ hours, all GPU had stopped mining with exit code 46. The miner just kept on going at 0sol/s.

I’m trying to figure out what might be causing these issues. I’ve searched around and some people with similar issues were able to fix or workaround by setting max power of each GPU to 94% out of 100%, maybe lower the OC speed ( my GPUs are sold already overclocked), or get new/better PCI-e risers.

I can try getting another set of risers, and that’ll take a few days to deliver. Anyone have thoughts or prayers? How would one go about overclocking or undervolting on Linux?

[root]

NVIDIA GTX 1080 ti GPUs

Here is a quick PowerShell (3) script to keep your EBWF miner running in the case that it fails and exits while you’re away. It’s pretty simply, the script checks for the process and if the process isn’t running it attempts to restart it every 5 minutes. You can change that value (Sleep 300) at the end to less or more. You’ll need to change the batpath, error path (or comment that out) and email information. This uses Gmail’s smtp and you will need to setup an application-specific password to …

Make sure you are using --eexit 3 in your .bat file.

[smokinggun46]

i’ll try this out soon.

btw, is your script based on windows? Or Linux?

[root]

I’m running it on Windows, but it’s a powershell script, so you could run it on windows but you might need to modify some parts of it.

[smokinggun46]

I appreciate the script.

The issue to my failing GPUs were the risers. They were bad risers, I’m not sure what component of the riser was causing the issue. Since my mobo had 2x 16x PCI-e, i plugged 2 cards into the mobo and they’ve been working perfect for the last 36 hours.