Messages window says new tasks requested, then gets these:
7/4/2009 10:53:54 PM Docking@Home Message from server: No work sent
7/4/2009 10:53:54 PM Docking@Home Message from server: (won't finish in time) BOINC runs 85.4% of time, computation enabled 99.9% of that
That happens sometimes when the workunits are very big. For the next batch of workunits we will increase the tim-out bound so that more machines can get jobs.
Thanks
ID:
5118 | Rating: 0
| rate:
/
Tom Harvie
Joined: Oct 4 08
Posts: 1
ID: 2064
Credit: 16,547
RAC: 0
I keep getting this message - Docking@Home Message from server: (won't finish in time) BOINC runs 95.5% of time, computation enabled 99.9% of that... what does this mean?
I keep getting this message - Docking@Home Message from server: (won't finish in time) BOINC runs 95.5% of time, computation enabled 99.9% of that... what does this mean?
It usually indicates that Boinc Manager has downloaded more work than it can handle, or at least more than it thinks it can handle.
When Boinc Manager adds up all the estimated run times of all currently downloaded work it believes that if it downloads more Docking then it wont finish it, even though you run your computer most of the time.
It could also be that the time for a work unit to run has been estimated incorrectly by the project and fools Boinc Manager into thinking they will run for longer that they will actually run.
Boinc will learn the correct run time as more work is processed and then will download more work.
____________
The host sends an invalid information to the server, the <resource_share_fraction> value. This is evaluated over all projects, including the disabled ones and including those that are set to "no new work".
So even if Docking is currently the only active project on that box, the client might report that it gets only 5% of the BOINC shares.
If you are sure that it will be able to do the results in time, it should help to either increase the ressource share until the cache is filled, or set the ressource shares of all inactive projects to 1 (and contact each of those projects).
______
From what I can see in the sources, a second irrelevant information might trigger this message too. The host sends the list of downloaded WUs (all projects) with their <cpu_time_remaining>. Basically cumulating those values can help to decide about potentially missed deadlines - but the list includes results from non-CPU-intensive projects, where the remaining time plays no role for any deadline estimation.