gaia@home

Posts by Keith Myers

1) Message boards : Number crunching : Feeder backend not running. Project in backoffs (Message 756)
Posted 13 days ago by Keith Myers
Post:
Until the admin makes a re-appearance here, the project is effectively dead, titsup, gone wobbly, whatever your favorite euphemism for exists no more.
2) Message boards : Number crunching : Feeder backend not running. Project in backoffs (Message 753)
Posted 16 days ago by Keith Myers
Post:
Feeder backend not running. Project in backoffs
3) Message boards : Number crunching : A lot of errors (Message 676)
Posted 8 Dec 2024 by Keith Myers
Post:
Wrapper script is misconfigured by the work generator. Benign error. Nuisance is only the extra cluttering of the stderr.txt output file.
4) Message boards : Number crunching : Ghost tasks for gaia 4 (Message 661)
Posted 5 Dec 2024 by Keith Myers
Post:
I've been getting Gaia_4 task all along, today included. They are misconfigured it appears but they validate with very low credit.
5) Message boards : Number crunching : 0 points for valdiated task ? (Message 550)
Posted 31 Oct 2024 by Keith Myers
Post:
I did click the first link. Can't comment on your issues since you choose to hide your hosts. It looks like all the 0 credit issues have to do with Gaia_4 tasks which were faulty and likely cancelled in bulk by the admins to make room for the Gaia_5 tasks.
6) Message boards : Number crunching : 0 points for valdiated task ? (Message 548)
Posted 31 Oct 2024 by Keith Myers
Post:
I have no idea why .....
First time is see this problem

I keep my fingers crossed for Ukraine

Seen this before on other projects. When there are already a _0 and _1 task already issued and with a deadline fast approaching, the scheduler preemptively issus a _2 task just in case neither of the two are expected to report in time. This occurs usually within an hour or two of the deadline.
When one of the first tasks reports and just squeeks in under the deadline, the _2 task is cancelled with a 0 credit assigned. This value then gets mistakenly applied to validated tasks.
Seen it often at Universe and now at Asteroids too. Looks like Gaia is susceptible also. It's a rare occurrence.
7) Message boards : Number crunching : Gaia 5 (Message 542)
Posted 29 Oct 2024 by Keith Myers
Post:
My one troublesome host has now been able to correctly finish and report these new 8560 GFLOPS tasks. Hoping we are finally in the stable part of the learning curve for the project.
8) Message boards : Number crunching : Gaia 5 (Message 538)
Posted 28 Oct 2024 by Keith Myers
Post:
I still can not get one slow host to successfully run the Gaia_5 tasks, even with this new 8560 GFLOPS size value.
9) Message boards : Number crunching : Long running Wu's ? (Message 537)
Posted 28 Oct 2024 by Keith Myers
Post:
Can you advise for tasks that hit 100% completion somewhere short of 3 hours and then just accumulate time until 3 1/2 hours have elapsed and then report as exceeded time limit errors.

I have one host that has not been successful in completing validated tasks since your original project resumption with Gaia_4 tasks.

Should I just give up on getting the tasks to properly run on this host?
10) Message boards : Number crunching : Gaia 5 (Message 524)
Posted 27 Oct 2024 by Keith Myers
Post:
Still cannot prevent one slow host from erroring out all its tasks. Going to NNT for now.
11) Message boards : Number crunching : Gaia 5 (Message 520)
Posted 27 Oct 2024 by Keith Myers
Post:
It could be some older work showing up... I have recently been getting 5_5xxxx work but also some 5_3xxxx resends mixed in at times.

I could have sworn they were new _0 tasks and not resends. The app speed is still too fast for this host.

Computer: Numbskull
Project: https://gaiaathome.eu/gaiaathome/ | Gaia@home
App Version: 1
App User: 5_Gaia@home
Application: 5_Gaia@home[20241023.2232]
Name: 5_57468
Name Wu: 5_57468_0
Elapsed: 01:01:17
Cpu: 99.00%
Progress: 100.000%
Timeleft: 00:00:00
Deadline: 9d,22:41:16 | Nov 05, 2024, 02:47:24 PM
Resources:
Status: Computation error
Exit status: 197
Platform: x86_64-pc-linux-gnu
Recieved: Oct 26, 2024, 03:47:25 PM
Disk: 953.6743 MByte
Task size: 39560.3310 GFLOPs (rsc_fpops_est)
App speed: 395603.3101 GFLOPs/sec (rsc_fpops_bound)
Memory: 476.8372 MByte (rsc_memory_bound)
12) Message boards : Number crunching : Gaia 5 (Message 516)
Posted 26 Oct 2024 by Keith Myers
Post:
I don't know why @zupa decided to change the values of the template again back to what was broken at the onset of the project return. The previous 39,560 GFLOPS speed was working just fine. Now tasks are back to erroring out with time exceeded limit errors.
13) Message boards : Number crunching : Gaia 5 (Message 514)
Posted 26 Oct 2024 by Keith Myers
Post:
It looks like Gaia tasks don't honor BOINC suspension. Boinc Manager says the tasks are suspended but the OS still shows running processes. This is under Linux

Are you sure?? Suspending tasks won't suspend the wrapper app which is just sleeping during suspend and actually during most of the crunching. The main app is suspended though and that is the one that actually does the crunching.
The wrapper app is the process with big 'G' in the name. The process with the little 'g' in the name is the actual science app.
14) Message boards : Number crunching : Gaia 5 (Message 512)
Posted 26 Oct 2024 by Keith Myers
Post:
Looks like the latest cutdown in the GFLOPS value for the tasks is working. Too bad it took them several days to properly configure their work generator.

Now out of work. Shame all those prior tasks were wasted.
15) Message boards : Number crunching : Gaia 5 (Message 509)
Posted 25 Oct 2024 by Keith Myers
Post:
I see the latest tasks have had their rsc_fpops_est values cut by 10X again. Now just 126,400 GFLOPS and much closer to the original 3,600 GFLOPS that was working for mostly everyone when the project first came back.
Hope to see this trend continue when we can expect all tasks be immune from the exceeded time limit errors.
[Edit] Still getting timeout errors on my slower Epyc's. They need to cut the GFLOPS values by another 10X.
16) Message boards : Number crunching : Gaia 5 (Message 507)
Posted 25 Oct 2024 by Keith Myers
Post:
I looked further and had some fail at the 15 minute 30 second point too. But the majority successfully validate at around 3.5 hours. So maybe just a bad series you got and other series are good.
17) Message boards : Number crunching : Gaia 5 (Message 505)
Posted 25 Oct 2024 by Keith Myers
Post:
@zupa Did you actually set up the wrapper to just crunch for 3.5 hours elapsed time on the tasks no matter their completion percentage? LIke how Rosetta does it for their tasks?
18) Message boards : Number crunching : Gaia 5 (Message 484)
Posted 24 Oct 2024 by Keith Myers
Post:
@Conan He could just kill all the 4s and recast them with the 5 settings, couldn't he?

Well, I just took another batch of Gaia 4s and looks like all 60+ of them are erroring out after running for 1h39m. I'm NNT till they're all gone from the que. Back to Asteroids for tonight...

+1
19) Message boards : Number crunching : Gaia 5 (Message 482)
Posted 23 Oct 2024 by Keith Myers
Post:
If he doesn't understand what he is tweaking in the task properties, he should ask for help in the Boinc developer forums or from other knowledgeable project admins. I can suggest the admins/scientists at GPUGrid who know the ins and out of the rsc_fpops_est values for the work generator templates quite well as one project also with very short task deadlines.

@Conan Appreciate you helping to remove all the bad tasks. My hosts have plenty of other cpu work they could be crunching with the wasted cycles on Gaia@home. So I set NNT this morning when I saw this issue was not going to be fixed quickly.
20) Message boards : Number crunching : Gaia 5 (Message 480)
Posted 23 Oct 2024 by Keith Myers
Post:
All 4_Gaia tasks still erroring out for exceeded time limits with the new task properties.

So why haven't you deleted all these 4_Gaia tasks which are pointless so that hosts can get the new 5_Gaia tasks that you say have been fixed with proper app speed changes?


Next 20

©2025 GAVIP-GC