Runnimg Seti@H, CPDN, LHC, and Pirates@H all set for 100 resource share, v 4.09 just crunches 1 hour then moves on to the next project, regardless of due date.


My name is Pascal and this message has my approval...

It is 10 oclock, do you know what your WUWUs are doing tonight...

Boinc also ignores the deadline period when asking for work: my pref were set to 7 days and I attach to LHC and I got far more wu's then what the computer could process in 2 days(deadline)even if LHC got 100% of the resources.

Due dates should NOT be a short time termination of results that

have simply not gotten back in time. Many real-life variables cause random

burps in "scheduled processor availability" Therefore time to return results

should be construed quite liberally by the project control mechanism

before reissue of the tardy WUs. I trust the seti project has such

somewhat liberal policy. Machines do go down for a time and are brought

back on line after troubleshooting/repair/return_from_vacation :-)....

WUs thus backlogged are then subsequently transmitted and should be credited.

Alfred

> Boinc also ignores the deadline period when asking for work: my pref were set
> to 7 days and I attach to LHC and I got far more wu's then what the computer
> could process in 2 days(deadline)even if LHC got 100% of the resources.

That short deadline was to clear the work units being processed before the new client was available. The old WU should be abandoned now so that you will get the newsest "Sixtrack" 4.45 and new work units.


<p>
For BOINC Documentaion: Click Me!

> Due dates should NOT be a short time termination of results that
>
> have simply not gotten back in time. Many real-life variables cause random
>
> burps in "scheduled processor availability" Therefore time to return results
>
> should be construed quite liberally by the project control mechanism
>
> before reissue of the tardy WUs. I trust the seti project has such
>
> somewhat liberal policy. Machines do go down for a time and are brought
>
> back on line after troubleshooting/repair/return_from_vacation :-)....
>
> WUs thus backlogged are then subsequently transmitted and should be credited.
>
> Alfred
>
Not the way that it works. If you get the work in before the replacement comes back, then you will get credit (so will the replacement). Otherwise, the result is move to offline storage, and you will not get credit. If you know that you are leaving on vacation in a week, set your cache to very short to minimize the number of WUs that you will need to drop when you get back.

> Runnimg Seti@H, CPDN, LHC, and Pirates@H all set for 100 resource share, v
> 4.09 just crunches 1 hour then moves on to the next project, regardless of due
> date.

All the 4.xx releases have been doing this.