Open Skilrok opened 3 years ago
Please try running your application like this:
CUDA_VISIBLE_DEVICES=-1 ./VerthashMiner --cl-devices-all (add more params here)
In this way you would mask your CUDA devices and would make them unavailable for the current solution. Hopefully that would work.
Hi,
I have an 8 gpu rig with 6 nvidia gpus, and 2 amd gpus. I am runing verthash miner 0.7.2 on HiveOS 0.6-203@210415. When I attempt to configure the miner to only utilize the amd gpus with the --cl-devices-all command, it starts 8 threads and utilizes all gpus. The closest I've gotten to success is --cu-devices 0 --cl-devices 0,1 and "sacrificed" one nvidia gpu to the miner.