I see your hours, not sure if they were there when you posted. It has a red color as if it was inactive.

160 climateprediction.net OpenIFS 43r3 Linear grid tl255 l91 3.40 3.40

Yeah I knew it was there I was just wishing I could get some of them, I have the Project asking for tasks at a high resource share but think since my cache is full of other things I'm not asking at the right time.

I have 280 threads running 4_Gaia and this has been consistent for at least a week now. 280 threads times 24 hours should be close to 6710 runtime hours per day but WU_Prop is only showing low 5000s. I would expect the hours to be within about 500 of 6710 depending on timing of WU reporting etc. I seem to be missing about 1500 hours everyday. I have always suspected the WU_Prop work units to miss some hours as they end, wait for upload, and then restart but not 1500 per day.

You have a lot of this in your task details where no data is being collected.

Erreur assignation taille wu
10:46:53 (446898): Erreur reception active_result

Nothing I have any control over. Nothing out out of the ordinary happening on any of the machines running the code. All WUs start and end normally as expected (including wu_prop units).

This is potentially a big part of the problem:

Tue 17 Dec 2024 09:06:59 AM CST | WUProp@Home | Scheduler request to https://wuprop.boinc-af.org/wuproj_cgi/cgi failed: Server returned nothing (no headers, no data)

Has happened several times recently on different systems and other members have reported it too. The only way to fix is to reset the project. Almost 9 hours of compute time missing from a 128 thread server.

You have a lot of this in your task details where no data is being collected.

Erreur assignation taille wu
10:46:53 (446898): Erreur reception active_result

I just checked my computers and every one of them is getting the same error messages, multiple times.
So it seems a bigger problem

Is there an application that is not playing ball?

e.g.

<core_client_version>7.17.0</core_client_version>
<![CDATA[
<stderr_txt>
06:24:02 (1795329): initialisation
Erreur assignation taille wu
Erreur assignation taille application
Erreur assignation taille wu
Erreur assignation taille application
Erreur assignation taille wu
07:24:03 (1795329): called boinc_finish(0)

Conan

---

I have those too in every WU, but the hours seem to count right... probably.

---

All my systems seem to be counting hours fairly well (within 100 hours in a 24 hour period for systems with 32 threads or fewer). It's the 128 thread system that is missing about 1000 hours everyday. I'm wondering if there are other large systems that are missing hours too. If so, WUProp stats are not accurate. There is nothing indicating that anything is wrong on the system itself as all work units are executing in the time expected and are validating at the server end consistently. It's frustrating to know you are executing work you are not being given credit for. About 10 days or so ago I was consistently being credited with over 6000 hours everyday but then it stopped.

All my systems seem to be counting hours fairly well (within 100 hours in a 24 hour period for systems with 32 threads or fewer). It's the 128 thread system that is missing about 1000 hours everyday. I'm wondering if there are other large systems that are missing hours too. If so, WUProp stats are not accurate. There is nothing indicating that anything is wrong on the system itself as all work units are executing in the time expected and are validating at the server end consistently. It's frustrating to know you are executing work you are not being given credit for. About 10 days or so ago I was consistently being credited with over 6000 hours everyday but then it stopped.

Did you change the apps you are running about 10 days ago? Or are you still running the same apps? I know I run ie the Prediction app for 24 hours and often get 0.03 hours for it.

No I haven't. I have been concentrating on 4_gaia for almost 10 days now. I'm just convinced that the WUProp app doesn't track runtime accurately for some reason. It misses minutes on some systems and hours on others. One just has to realize that if you are trying to get 100,000 hours on an application you will have to run close to 125,000 hours to achieve it. You just have to factor in the lost 25% in your planning. Since the website specifically states runtime and not CPU time, I'm expecting a 32 thread machine to have 32 hours of runtime per wall clock hour. If I get 30.5 to 31.5 then I accept that as accurate but that isn't what I'm getting. I even selected "Computers on this account" and added the daily activity for each computer and it doesn't match the hours displayed on the ACCOUNT page. Like I said, I have accepted the fact that this site isn't accurate and I am factoring that in my planning.

No I haven't. I have been concentrating on 4_gaia for almost 10 days now. I'm just convinced that the WUProp app doesn't track runtime accurately for some reason. It misses minutes on some systems and hours on others. One just has to realize that if you are trying to get 100,000 hours on an application you will have to run close to 125,000 hours to achieve it. You just have to factor in the lost 25% in your planning. Since the website specifically states runtime and not CPU time, I'm expecting a 32 thread machine to have 32 hours of runtime per wall clock hour. If I get 30.5 to 31.5 then I accept that as accurate but that isn't what I'm getting. I even selected "Computers on this account" and added the daily activity for each computer and it doesn't match the hours displayed on the ACCOUNT page. Like I said, I have accepted the fact that this site isn't accurate and I am factoring that in my planning.

There is a small amount of time switching from one task to the next task and winding down of the one task and the winding up of the next task, etc, etc, etc. I know PrimeGrid had had discussions about this in the past and they have talked about it taking over 1 minute to happen. BUT I DO agree with you that that alone should NOT count in the time you are missing.

I am running "All-Sky Gravitational Wave search on O3" at Einstein, but none of the hours are showing up. Is anyone else having this same problem?

---

Reno, NV
Team: SETI.USA

I had it before, now they showed up.

---

And of course, as soon as I post about it, it is working again.

All good now.

---

Reno, NV
Team: SETI.USA

WUProp not picking up hours from the BUDA application from the LLMentorGrid project. This is an application running in a docker container so WUProp probably doesn't recognize it.

WUProp not picking up hours from the BUDA application from the LLMentorGrid project. This is an application running in a docker container so WUProp probably doesn't recognize it.

Do you have the website for the LLMentorGrid project?

Never mind I found it, Linux only though:
http://boinc.llmentor.org/LLMentorGrid/

WUProp not picking up hours from the BUDA application from the LLMentorGrid project. This is an application running in a docker container so WUProp probably doesn't recognize it.

It's now being recognized per this automatically generated post:

The application BUDA from project LLMentorGrid is active

I've been running it for almost 12 hours and it isn't showing up on my account. Over the last 3 to 4 hours, NO HOURS are being logged to my account. ALL the runtime hours are the same totals they were around 20:00 UTC.

SOMETHING IS BROKEN AGAIN

I'm seeing the same thing happening on all my active tasks - NO NEW HOURS BEING REPORTED .

PrimeGrid	Genefer 17 Mega	9,388.72		89.98

That app should be reporting at least 133 to 144 hours for last 24.

My hours now look OK, were low.

---

Paul.