the time to completion is not real.
first, this dont let boinc to download more than one work task per time
so i need to connect very often to report and download tasks
how can a make , boinc, to learn that a work unit takes 6 hours and not 300
usualy the time to completion starts at 300-500 hours and then decreases rapidly
there is a way to reset this at boinc
thanks

Which project are you talking about?

i have the seti project, eistein and roseta
but the problem is general
when the work task starts , has a 300-900 hours as time to complition, but then this time decreases , and at the end it takes only 4-6 hours per task
the problen is that i dont get on line every day , and a lot of work task is lost ( i can make more if the boinc understand that the time is not so high)
i put the preferencer to connect every 10 days
the first month, every thing was ok , and boinc download enought work for a week
any clues ?


thanks

It says 300-900 hours for all projects? That's really strange.

Try this:
Stop BOINC. Don't just close BOINC Manager, that will keep it on tray, use File | Exit.
Open client_state.xml (on BOINC installation directory) with a text editor. Notepad will do just fine.
Search for <duration_correction_factor> (should be there once per project) and let me know what number it has, for each project. Don't change anything (and if you do, don't save the file).
If too low or too high, that would be the cause (normal should be around 1.0). If it seems normal, maybe you have a problem with benchmarks, or maybe the project is sending a bad time estimation...

i have
<duration_correction_factor>126.404841</duration_correction_factor>


should i change it to

<duration_correction_factor>1.0</duration_correction_factor>

thanks

ps: i did not find a search option to the forum. there is ?

Yes, you should. Remember to do it with BOINC not running.

Thanks . it works . great help

thanks pov, i had same problem, but chose to ignore it. Fixed it now.

BBC now reads 4000 hours to finish not 280000+ hours

Eek, that's big!

Somebody had a similar problem on my project, see this forum thread

My fastest box has a simular problem but unfortunaly it is not duration correction factor nor is it the benchmarks. Any other suggestions?

BoB