I couldn't find this when I searched for it. The BOINC client has not updated the "work done for this project" in a long time, and I have restarted my Windows machine at least once since the symptom showed up. It is actively getting work to do, and is regularly updating the about of work it has completed on each work chunk.
Copyright © 2024 Einstein@Home. All rights reserved.
KR Fleming wrote: I couldn't
)
Be sure you have YES checked on this page to export your stats:
https://einsteinathome.org/account/prefs/privacy
KR Fleming wrote:... The
)
Where exactly are you looking to see this information?
If you look at your account page, you will see a list of your computers and the current statistics for each one. Are you saying that those numbers haven't changed "in a long time"?
If you are looking at an external statistics site that records details for all sorts of BOINC projects, you won't be able to see any details from your account if you haven't given permission for Einstein@Home to export statistics to these sites. It's a privacy requirement for you to 'allow' such exports.
I can see your current stats under your name in the margin of your post, but because your computers are 'hidden' I can't see any details about individual machines on your account. For that reason, there is no way for me to check if there is some sort of problem. You need to change your preferences (as suggested by Mikey) if you wish for others to be able to check. Only non-sensitive details are shown so it's quite OK to do that.
Cheers,
Gary.
Thanks Gary. I'm looking at
)
Thanks Gary. I'm looking at the BOINC Manager (7.22.2) as well as my Account Page here at this site. Selecting YES on the Export Stats option per Mikey's recommendation didn't change anything--I'd estimate that the 'Work Done' statistic updated regularly for about the first 2 weeks or so of processing, up to the point it reached 2,389,106. It's been at that figure now for about a month, roughly, even though Status is usually Running. This isnt a crisis--everything else indicates work really is getting done, it's just a weird satisfaction I got seeing the little odometer turning.
You have a ton of errors and
)
You have a ton of errors and only a few valid tasks on your tasks page. I recommend going to project preferences and un-selecting Gravitational Wave work units. Also remember to select "no" to allowing tasks you didn't select otherwise it will just send them anyway and they will keep failing.
I do see that most if not all
)
I do see that most if not all errors were from the work units you suggested unselecting, which I've done. I did then opt to Reset Project because it sounded like a good idea, and returned to preferences to find the Gravitation Wave stuff was still de-selected. After all this, the first thing to start crunching was the All-Sky gravitational wave work :). But, looks like work has shifted over to other work units now. Will wait for something to successfully complete and check Work Completed again. Thanks!
KR Fleming wrote:... This
)
Actually it was a crisis since the GW GPU work being returned was not valid :-).
Thanks for 'un-hiding' your computers. We can now see your GPU has 4GB VRAM. Unfortunately this is insufficient for GW GPU tasks.
When new searches are started (like the current GW GPU search) there will be an announcement thread in the Tech News forum. If you are running a seach whose tasks seem to be having problems, it's always a good idea to check to see if others are reporting the same issue in the announcement thread.
Your problem is exactly the same as the one in this report which was posted back on July 12. That user was commenting on the memory requirements and the following comment confirmed that 4GB VRAM was indeed insufficient for these tasks.
You can also check this for your own task failures by finding your computer on the website and clicking the tasks link to get a full list of all the tasks your computer is running. I clicked on the task ID link for this failed task to see what was returned to the project. If you use that link you will see exactly what I saw. Scroll down below the Stderr Output sub-heading, looking for words like 'failed' or 'error' etc. Ignore warnings. You will find this famous line:-
Enqueue OpenCL kernel failed with OpenCL error: CL_MEM_OBJECT_ALLOCATION_FAILURE
which tells you exactly what the problem was - a failure to allocate sufficient memory needed to run the task.
You can diagnose lots of problems this way by developing the habit of following the task ID link for tasks that happen to fail.
Cheers,
Gary.