no work available
Message boards :
Number crunching :
no work available
Message board moderation
Author | Message |
---|---|
Send message Joined: 3 Feb 24 Posts: 3 Credit: 0 RAC: 0 |
Why is no work currently available for denis@home? |
Send message Joined: 8 Jul 22 Posts: 9 Credit: 964,279 RAC: 0 |
See below news item, may be some soon. https://denis.usj.es/denisathome/forum_thread.php?id=300 Paul. |
Send message Joined: 6 Mar 23 Posts: 40 Credit: 2,078,354 RAC: 0 |
See below news item, may be some soon. He suggested there would be new tasks "this afternoon." That was July 28. There have breen none since. |
Send message Joined: 8 Jul 22 Posts: 9 Credit: 964,279 RAC: 0 |
He suggested there would be new tasks "this afternoon." That was July 28. That thread only started July 29th, last post August 20th said preparing, no date given. https://denis.usj.es/denisathome/forum_thread.php?id=300&postid=2466 Paul. |
Send message Joined: 9 Apr 15 Posts: 173 Credit: 1,552,856 RAC: 0 |
That thread only started July 29th, last post August 20th said preparing, no date given. Maybe for Christmas... |
Send message Joined: 3 Feb 24 Posts: 3 Credit: 0 RAC: 0 |
It’s now 2025, and still no WU’s available. |
Send message Joined: 8 Apr 15 Posts: 35 Credit: 389,238 RAC: 0 |
It’s now 2025, and still no WU’s available. Don't get your hopes up for work anytime soon with this project |
Send message Joined: 6 Mar 23 Posts: 40 Credit: 2,078,354 RAC: 0 |
Don't get your hopes up for work anytime soon with this project Well, no work from ClimatePrediction since about last November, and even then, only a very few Windows tasks. Nothing since last August for my Linux machine, and they were just re-runs. No work from WCG all month. They are supposed to start again today, but no sign of them yet. As a practical matter, they have been mostly down for about two years, though there have been spurts of MCM1 tasks once en a while and I got bunch of ARP1 tasks that completed successfully, but there has been no trace on their web site. No work from Denis since about last July, Now, Rosetta has stopped sending my machines any work.. So I have been running 12 Einstein tasks at a time even though it is the project I am least interested in. It makes me think that Boinc style work is becoming obsolete. I do not know if it is due to Boinc being too slow and unreliable, new supercomputers becoming too cheap, or what. |
Send message Joined: 29 Sep 24 Posts: 1 Credit: 0 RAC: 0 |
Don't get your hopes up for work anytime soon with this project WCG has been fully operational for at least several months, and with the exception of a few weeks of poor performance when ARP1 was restarted, had been consistently handing our WUs. They provided advance notice that they would be down through January 3rd (which I interpreted to include the 3rd, so I don't expect them to return to operation until next week). While WCG was was down, they reported making progress on a new project -- Mapping Arthritis Markers -- and also discussed things they were doing to potentially enable MCM to run on GPUs. You can see the details of this at https://www.cs.toronto.edu/~juris/jlab/wcg.html (click on the "Operational Status" tab). Some projects (e.g., Asteroids@Home, LHC@Home) simply ran out of WUs for people to crunch, and since it happened over the holidays, the folks who generate these WUs were likely out for the year. If that's all there is to it, I'd expect the situation to resolve over the next 1-2 weeks as people return from their holiday breaks. Rosetta in particular has been disrupted by fundamental changes in its niche due to advances in AI that has significantly changed the computational needs of this type of work (presumably in a way that isn't amenable to running in a distributed computing project -- at least with currently-available consumer hardware). Not sure what the deal is with other projects (including DENIS), but whatever the case, it's likely a result of that particular project's situation, rather than BOINC as a whole becoming obsolete. One area where BOINC is suffering from technical limitations (and was explicitly called out by WCG during their winter downtime announcements) is that scientific computing workloads are increasingly expected to be run in a Docker container for convenience, but BOINC doesn't have an easy/native way to run such workloads. There is ongoing work on adding Docker support to BOINC (under the name "BUDA"), and it's supposed to be rolled out with the next release. There's further discussions on this point on the last few months' posts on a blog maintained by one of the BOINC developers: https://aenbleidd.blogspot.com/2025/01/boinc-development-status-report.html Overall, I think it's premature to write off BOINC (and distributed computing in general) as being outdated. BOINC was always designed to run multiple projects concurrently, and I've got multiple projects active at the moment, with preference given to biological science projects (including DENIS, if and when it returns with work to do). This keeps my machines continually busy as the work for various projects comes and goes. And if you're mainly interested in biology work, Folding@Home -- while not a BOINC project -- is still very much active. |