Recently upgraded my most problematic one from 1.4.3 OpenCL to metashape 15.1 CUDA and sure enough got the same error immediately on starting alignment(5069 16 MP images in two camera groups):Įrror: CUDA_ERROR_UNKNOWN (999) at line 128 Thanks that was the problem, I had worked around it by setting the nvidia device permissions to 666 but this is better.I have been having CUDA errors on a couple of our workstations since v. Posted: Mon 2:29 pm Post subject: Re: Docker nvidia GPU errors #debug = "/var/log/nvidia-container-runtime.log #debug = "/var/log/nvidia-container-toolkit.log" May be you should edit the /etc/nvidia-container-runtime/config.toml,Īnd add user = "root:video" to fix this problem. Linux ryzen 5.6.17-gentoo #1 SMP PREEMPT Mon Jun 8 11:35: x86_64 AMD Ryzen 9 3900X 12-Core Processor AuthenticAMD GNU/Linux | Fan Temp Perf Pwr:Usage/Cap| Memory-Usage | GPU-Util Compute M. | GPU Name Persistence-M| Bus-Id Disp.A | Volatile Uncorr. $ docker run -gpus all nvidia/cuda:10.0-base nvidia-smi nvidia-docker2 ) issue under gentoo and got this working.Īssume you have add yourself in docker groupĭocker run -gpus all nvidia/cuda:10.0-base nvidia-smi I fixed the nvidia-container-toolkit (aka.
Posted: Fri 5:39 pm Post subject: Re: Docker nvidia GPU errors Last edited by AlexJGreen on Mon 3:28 am edited 1 time in total
ID: Q6EP:ZENB:6ZUS:JZBI:3IZX:VGMS:YQCD:JPJY:PI6L:AN3G:PFVW:54OY Log: awslogs fluentd gcplogs gelf journald json-file local logentries splunk syslogĬontainerd version: 35bd7a5f69c13e1563af8a93431411cd9ecf5021 Network: bridge host ipvlan macvlan null overlay If I try to chcon the devices it gives an ~ $ sudo chcon -t container_file_t /dev/nvidia*Ĭhcon: failed to get security context of '/dev/nvidia0': Operation not supportedĬhcon: failed to get security context of '/dev/nvidiactl': Operation not supportedĬhcon: failed to get security context of '/dev/nvidia-modeset': Operation not supportedĬhcon: failed to get security context of '/dev/nvidia-uvm': Operation not supportedĬhcon: failed to get security context of '/dev/nvidia-uvm-tools': Operation not supported I'm trying to get nvidia gpus to work inside docker.īut when I try to run it gives an ~ $ docker run -rm -gpus all nvidia/cuda:9.0-base nvidia-smiĭocker: Error response from daemon: OCI runtime create failed: container_linux.go:349: starting container process caused "process_linux.go:449: container init caused \"process_linux.go:432: running prestart hook 0 caused \\\"error running hook: exit status 1, stdout:, stderr: nvidia-container-cli: initialization error: nvml error: insufficient permissions\\Īfter reading, it looks that it might be an SELinux issue but I'm not running SELinux, although I have docker compiled with seccomp enabled. Posted: Wed 4:34 pm Post subject: Docker nvidia GPU errors
#Cuda driver version is insufficient for cuda runtime version mac os x 10.13.4 software#
Gentoo Forums Forum Index Unsupported Software Gentoo Forums :: View topic - Docker nvidia GPU errors
0 kommentar(er)
