The server status says that there are 810124 tasks ready to send, but I'm not getting anything sent to my BOINC client to work on them, even if I update the project.

Any ideas why this may be happening? I appreciate any insight into this.

~MasterMind

I also seemed to have that problem, but tasks started downloading once I rebooted my computer.

Greetings !!
Message removed. and replaced.

About an hour ago, I powered up an started BOINC.
Initially, A@H said no WU were available.
Then, BOINC retried for about 40 minutes at various intervals.

Then, Wu were downloaded.

is waiting for a few tries now the new normal?

Thanks !!!
Jay

Seems like there's a scheduler problem, that sometimes stops sending work out. I've been crunching mostly pretty steady for a couple weeks on my Nvidia GTX 980, and sometimes a few of my systems get CPU jobs. But today it reports no tasks available all day, for the computer with the GTX 980, both for GPU and CPU. Server status page says there are almost 6 million tasks ready to send right now, and that number has been steadily increasing over the past couple weeks. Tested forcing Asteroids update and suspending other projects on a different PC, it got a CPU task first try. Rebooted, the GTX 980 computer, but still nothing so far.

I've found a few other threads scattered around the forum talking about either suddenly no tasks, or waiting a while and continuously retrying eventually gets new tasks. But today, waiting doesn't seem to be working...

Edit: Aha! got some tasks now, after I did a force Asteroids update on another computer of mine. It got 49 CPU tasks right away, and now that I've checked back to this computer, it's gotten the GPU tasks it's been requesting all day. Maybe this will help someone else, while the project admins work out their end.

je possede une rtx 4000 sff ada et une rtx pro 2000 blackwell dans le meme pc.
j'ai trompé le serveur en formatant mon pc windows 10 et en gardant juste la rtx 4000 sff ada a l'intérieur.Cela m'a permis d'avoir 100 unités a calculer.Par contre,des que j'ai remis la rtx pro 2000 blackwell dans ma tour,le serveur a détecté 2 rtx pro blackwell et ne m'envoie plus de calcul.J'ai lancé un script automatique pour demander du travail toutes les 15 mn mais cela ne change rien.J'empeche astéroids@home de calculer sur la rtx pro blackwell en passant par le fichier cconfig.file de boinc mais en vain.il faudrait mettre a jour le serveur vers les gpu blackwell et revoir le systeme de détection des cartes graphique.J'espere que la mise a jour vers les gpu blackwell ne va pas tarder car les unités de calcul s'accumulent rapidement.

I have an rtx 4000 sff ada and a rtx pro 2000 blackwell in the same pc.
I tricked the server by formatting my windows 10 pc and just keeping the rtx 4000 sff ada inside. This allowed me to have 100 units to calculate. On the other hand, as soon as I put the rtx pro 2000 blackwell back in my tower, the server detected 2 rtx pro blackwell and no longer sends me a calculation. I started an automatic script to request work every 15 minutes but that doesn’t change anything. I prevent astéroids@home from calculating on the rtx pro blackwell through the boinc cconfig.file but in vain. It would be necessary to update the server to the blackwell GPUs and review the graphics card detection system. I hope that the update to Blackwell GPUs won’t take long because the compute units are accumulating quickly.

---

Since it appears that the project is ignoring your Blackwell gpus exclusion in cc_config.xml if I understand your statement, I would go to the option of removing the detection report in the coproc_info.xml file and lock the file down as immutable so the client can't rewrite it anymore.

---

A proud member of the OFA (Old Farts Association)

bonsoir et merci.Je viens d'essayer et j'ai bien 2 rtx 4000 sff ada qui s'affiche dans boinc mais cela ne fonctionne pas car je n'ai pas eu de travail malgré le script régulier.je vais attendre .Je suis déçu car j'ai l'impression que boinc devient une brocante(gpugrid puis asteroids@home).Personnellement, je n'aime pas trop les bidouilles en informatique pour que les choses fonctionnent.

good evening and thank you. I just tried and I do have 2 rtx 4000 sff ada displayed in boinc but it doesn’t work because I didn’t get any work despite the regular script. I will wait. I am disappointed because I have the impression that boinc is becoming a flea market (gpugrid then asteroids@home). Personally, I don’t really like the hacks in computing for things to work.

---

bonjour
finalement,j'ai retiré ma rtx 2000 pro blackwell pour la remplacer par une rtx a 2000 et cela fonctionne.Je reçois des unités de travail.Je vais attendre la mise a jour vers les gpu blackwell pour la remonter.

hello
finally, I removed my rtx 2000 pro blackwell to replace it with an rtx at 2000 and it works. I am receiving work units. I will wait for the update to the Blackwell GPUs to bring it up.

---

Bonjour,
Il est vrai que c'est pénible de devoir "chipoter" pour obtenir des UT.
Pour GPU ça va encore, mais pour CPU ...
Du coup ça donne juste envie de zapper le projet.
Bonnes fêtes,

I heard there is some DoS prevention script running on this project's server which not only prevents hammering the systems but also prevents fast machines from getting work...

None of my machines can get work sincs months.

A few days ago, I tried Ryzen 5950X and Threadripper Pro 9000 machines using RTX 3080, Blackwell 6000 Pro and AMD 9700 Pro AI GPUs. Not a single task received using the latest BOINC client (and also tested the older 8.0.2 version).

If the project team is interested in fast machines, it would be helpful if they could take the time and solve the problem.

Michael.

---

President of Rechenkraft.net e.V.

je pense qu il faut retirer les gpu blackwell car,apparemment,le projet n'est pas compatible.Lorsque je monte ma rtx 2000 pro blackwell,je n'ai pas de travail et si je la retire,je reçois jusqu'a 100 unités.les gérants du projets n'ont pas l'air de s'affoler alors moi non plus.J'ai remis ma rtx 2000 blackwell,terminer mes 100 unités avec ma rtx 4000 sff ada, et attendre une évolution.

I think that it is necessary to remove the blackwell GPUs because, apparently, the project is not compatible. When I install my rtx 2000 pro blackwell, I have no work and if I remove it, I receive up to 100 units. The managers of the projects do not seem to be panicking so neither do I. I put back my blackwell RTX 2000, finished my 100 units with my RTX 4000 sff ada, and waited for an evolution

---

I have been getting this error message for a couple of weeks now.

Thu Jan 1 13:52:45 2026 | Asteroids@home | Scheduler list fetch from https://asteroidsathome.net/boinc/ failed: transient HTTP error

I can see according to the server status page that there are work units available so why I am getting this error.

I am running in an iMac Pro 8 core Xeon W Radeon Vega 56 system.

Thanks for any assistance anyone can provide.

je pense qu il faut retirer les gpu blackwell car,apparemment,le projet n'est pas compatible.Lorsque je monte ma rtx 2000 pro blackwell,je n'ai pas de travail et si je la retire,je reçois jusqu'a 100 unités.les gérants du projets n'ont pas l'air de s'affoler alors moi non plus.J'ai remis ma rtx 2000 blackwell,terminer mes 100 unités avec ma rtx 4000 sff ada, et attendre une évolution.

I think that it is necessary to remove the blackwell GPUs because, apparently, the project is not compatible. When I install my rtx 2000 pro blackwell, I have no work and if I remove it, I receive up to 100 units. The managers of the projects do not seem to be panicking so neither do I. I put back my blackwell RTX 2000, finished my 100 units with my RTX 4000 sff ada, and waited for an evolution

I was talking about CPU tasks, too. Running Asteroids@home on GPU does not really make much sense in my opinion.

Hence, the problem I reported is in no way related to the GPUs I have installed.
Yesterday, to my surprise, both machines fetched a few tasks, completeted and returned these successfully. And then never got any additional work, although all WUs were validated.
At the same time, the project's server status reports thousands of available tasks...

So, there is some sever issue on the BOINC server side.
If you like the tasks to be completed - fix it.
Otherwise people move on to other projects as these these type of machines consume too much electricity in idle mode to leave the cores unoccupied.

Michael.

---

President of Rechenkraft.net e.V.