2.5 KiB
Render Group and Rebooting
When trying to do a hardware shutdown, I got these error messages for every instance
of the clinfo process that I had started up. They hang hard, and don't respond to
kill signals.
systemd-shutdown[1]: Waiting for process: clinfo, clinfo, clinfo, clinfo, clinfo
Rebooting
New groups indeed show up upon reboot.
$ id
uid=1000(arcologos) gid=1000(arcologos) groups=1000(arcologos),4(adm),24(cdrom),27(sudo),30(dip),44(video),46(plugdev),109(render),120(lpadmin),133(lxd),134(sambashare)
Killing clinfo Process
Repeatedly calling
ps -eax
and
sudo kill -KILL <pid>
of anything mentioning clinfo eventually did kill the bash processes
waiting on clinfo after many minutes. However, the real processes
remain as zombies
$ ps -eax | grep clinfo
1519 pts/0 D 0:44 [clinfo]
1553 pts/1 D 0:00 [clinfo]
Installing libnuma-dev
One of the GitHub issue advice was to install libnuma-dev,
and now I wish I had been paying more attention to NUMA when
working on SaLSa self-driving cars with Samhitha.
$ sudo apt install numactl
Reading package lists... Done
Building dependency tree
Reading state information... Done
The following NEW packages will be installed:
numactl
0 upgraded, 1 newly installed, 0 to remove and 171 not upgraded.
Need to get 38.5 kB of archives.
After this operation, 150 kB of additional disk space will be used.
Get:1 http://us.archive.ubuntu.com/ubuntu focal/main amd64 numactl amd64 2.0.12-1 [38.5 kB]
Fetched 38.5 kB in 0s (140 kB/s)
Selecting previously unselected package numactl.
(Reading database ... 165065 files and directories currently installed.)
Preparing to unpack .../numactl_2.0.12-1_amd64.deb ...
Unpacking numactl (2.0.12-1) ...
Setting up numactl (2.0.12-1) ...
Processing triggers for man-db (2.9.1-1) ...
The two cores of the Celeron show up in numactl but not the GPU.
$ sudo numactl -s
policy: default
preferred node: current
physcpubind: 0 1
cpubind: 0
nodebind: 0
membind: 0
$ sudo numactl --hardware
available: 1 nodes (0)
node 0 cpus: 0 1
node 0 size: 1574 MB
node 0 free: 103 MB
node distances:
node 0
0: 10
Perhaps a Problem with rocm-opencl or Ubuntu distribution.
A problem with this on Fedora prompts clinfo maintainer to suggest taking it up with Fedora
https://github.com/Oblomov/clinfo/issues/81
Unanswered https://community.amd.com/t5/drivers-software/clinfo-gets-hanged/td-p/444906