Error while computing
7 Nov 2016 19:08:08 UTC
Topic 202609
(moderation:
Hello guys.
I am (was) running 2 CPU tasks at the same time. But then i decide to stop one (suspend), because i am going to need some more CPU power while gaming. I suspend it. After ~20 seconds, i go to the Task Manager and kill the CPU task process, because 500MB of RAM are not a small amount for something that is not running. Aaaand i get right away Error while Computing. And i remember doing it 3-4 times before that without any issue. I checked the Task exit code : Unknown error code 0x00000001. Any ideas why is this happening?
It seems better to just
)
It seems better to just change BOINC Manager to 50% CPUs instead of killing an executable.
There is an option to leave data in memory if paused/suspended. I forget the actual term of the option. Is that enabled leaving 500mb used when the task has been paused?
Ugh, yea, thats right, i have
)
Ugh, yea, thats right, i have Leave non-GPU tasks in memory while suspended?:
no
Enabled. Although for some reason, when i suspend a task, wait 1-2 minutes and start it again, it still starts with some 0.02% data loss. Checkpoints are every 60 seconds, so i may as well disable this option.
0.02% seems very ok. Other
)
0.02% seems very ok. Other projects checkpoint much less often. There was one I tried running on my tablet that I bring to work that never checkpointed so when I traveled to work and back I basically did nothing at all as WUs were longer than my work time.
Wait, doesn't specifically
)
Wait, doesn't specifically saying "MAKE CHECKPOINTS EVERY X TIME" works for everything?
Well, the option/preference
)
Well, the option/preference is called "Request tasks to checkpoint at most every: XX seconds" and the key words are "at most" so tasks will not checkpoint more often than specified. I reality there might be several minutes between checkpoints and some applications at other projects don't checkpoint at all.
I remember that some of the previous runs/apps here at Einstein could take up to 30 minutes between checkpoints on a fairly powerful machine.
Yyyeah, had a power failure
)
Yyyeah, had a power failure and lost 10% progress on the FGRPSSE task :/ Either have to blame the SSD or the amount of check points it created for that ~2 hours of work.