WU still exiting

[lamy]

HI

I installed a new nvidia 5070ti and did a clean install of mint linux 22.1.

it completed its first WU just fine but...

22:12:58:I1:WU2:Completed 1250000 out of 1250000 steps (100%)
22:12:58:I1:WU2:Average performance: 69.6774 ns/day
22:13:01:I1:WU2:Checkpoint completed at step 1250000
22:13:13:I1:WU2:Saving result file ../logfile_01.txt
22:13:13:I1:WU2:Saving result file checkpointIntegrator.xml
22:13:13:I1:WU2:Saving result file checkpointState.xml.bz2
22:13:13:I1:WU2:Saving result file positions.xtc
22:13:13:I1:WU2:Saving result file science.log
22:13:13:I1:WU2:Saving result file xtcAtoms.csv.bz2
22:13:13:I1:WU2:Folding@home Core Shutdown: FINISHED_UNIT
08:31:56:I1:WU2:Caught signal SIGINT(2) on PID 1240
08:31:56:I1:WU2:Exiting, please wait. . .

its stuck on exiting

I've not been able o find a comparable problem thus far to refer to.

I'm currently on 2.522 billion points

any ideas?

Lamy

[arisu]

How long has it been stuck on exiting? Does your process monitor show that it is using any CPU?

Try sending it SIGINT again. This might be a transient error.

[lamy]

Curiously the CPU based WU are fine it's something in the GPU units that will not finalise and send:
Work Unit #1
Status
Running
Progress
100.0%
Machine
RYZENDESK
OS
linux
TPF
6m 30s
PPD
4,883,946
CPUs
1
GPUs
NVIDIA GeForce RTX 5070 Ti
Run Time
10h 51m
ETA
0s
Assign Time
19h 12m ago
Base Credit
143,000
Deadline

WU has done the work: 11:16:19:I1:WU1:Folding@home Core Shutdown: FINISHED_UNIT
and that is as far as the client has gotten with the GPU based WU

Recent Work Unit History
Machine
Project
Core
OS
Status
Progress
TPF
PPD
Assign Time
Info
RYZENDESK
18803
0xa9
55.0%
1m 56s
58,369
12s ago
RYZENDESK
12468
0xa8
100.0%
53s
114,484
1h 28m ago
RYZENDESK
18240
0xa8
100.0%
2m 09s
91,579
5h 03m ago
RYZENDESK
18240
0xa8
100.0%
2m 26s
75,743
9h 08m ago
RYZENDESK
18230
0x24
100.0%
5m 29s
4,882,334
17h 30m ago


CPU based WU are fine and complete normally

I had been folding on an RTX2070 without issue. Am running mint 22.1

Lamy

[arisu]

Did you suspend (sleep) the system at all while it was folding? Sometimes that can cause the GPU driver to get stuck which prevents the core from being released. That's probably not the situation here unless you set it to sleep right as it finished, but it's a guess.

Do they still stall even after a reboot?

[lamy]

HI

I leave the pc on 24/7 it never sleeps.

Is there a FAH beta beyond 8.4.9 that might be available to test out newer hardware under linux?

Lamy

[muziqaz]

Nothing beyond current version.

[arisu]

There is the latest https://github.com/foldingathome/fah-client-bastet that you can compile yourself which has newer code than the 8.4.9 tag, but it is untested and doesn't contain any code changes that would be relevant to the problem you're having anyway.

[calxalot]

Someone reported problems with the bleeding edge build not connecting to servers.

I think some websocket api changes are not yet implemented in the node server.

Edit: I seems to be user1969 on discord that was building latest.