Folding Forum

Hello,

i cant get my GTX1060 to work. Here is my config:

19:19:15:<config>
19:19:15: <!-- Folding Core -->
19:19:15: <checkpoint v='3'/>
19:19:15: <core-priority v='low'/>
19:19:15:
19:19:15: <!-- Logging -->
19:19:15: <verbosity v='5'/>
19:19:15:
19:19:15: <!-- Network -->
19:19:15: <proxy v=':8080'/>
19:19:15:
19:19:15: <!-- Slot Control -->
19:19:15: <pause-on-battery v='false'/>
19:19:15: <power v='full'/>
19:19:15:
19:19:15: <!-- User Information -->
19:19:15: <team v='235850'/>
19:19:15: <user v='InstructionCacheLvl0FetchError'/>
19:19:15:
19:19:15: <!-- Folding Slots -->
19:19:15: <slot id='0' type='GPU'>
19:19:15: <cuda-index v='1'/>
19:19:15: <gpu-index v='1'/>
19:19:15: <opencl-index v='1'/>
19:19:15: </slot>
19:19:15: <slot id='1' type='CPU'>
19:19:15: <cpus v='8'/>
19:19:15: </slot>
19:19:15:</config>

I have a windows in a kvm virtual Machine on Archlinux running. In the VM i have a Windows 1809 running with normal Nvidia Drivers. I can play "Tripple-A" Titles just fine and can see usage of my GPU in Windows Taskmanager.

FAHClient ist just using one CPU Core instead of my GPU. Can someone pinpoint me to search for errors in Log? There arent any, i have already set the log to level 5.

Greetings from Germany

Are you getting "No WUs available for this configuration"? If so, it is because parts of F@H are overloaded with the enormous response to several social media references. The team are working on adding additional resources, please be patient. Just consider what it's like for the researchers who are generating many times more work than a week ago.

19:19:15: <verbosity v='5'/> after you set this back to 3, send us the first 200 lines of the log. They tell us so much.

The CPU count is set to 8, but your CPU is just using one. That might be from a installer bug that sometimes causes the first WU to be processed an just one CPU core. Later WU's will use more. Of course I am assuming your system has 8 or more CPU cores available to it.

Not sure about the rest of your setup, the client has problems with some type of VM's and accessing the GPU.

Actual CPU performance cannot be increased by creating virtual CPUs in VMs

Say you have 4 real CPUs and 2 VMs, each with 8 CPUs. Whether you have a client running in each machine that allocates N+M GPUs, you won't get any more work done than the 4 CPUs can produce, you'll get less because of the VM overhead.