Okay, everything is done. The ones that were returned late have validated; some of them have their extra tasks marked "didn't need" so they aren't causing anyone to do unnecessary work.
One of the ones I aborted still hasn't been sent to anyone else; if I hadn't aborted it, I would have finished it by now. Oh well. Maybe Boinc needs an "ask server for deadline extension" button.
Anyway, I will check into updating my video driver after work today. But since the span from the last bsod to this one was nearly four months, it may be a while before I know if the update helped.
David
Miserable old git
Patiently waiting for the asteroid with my name on it.
Okay, everything is done. The ones that were returned late have validated; some of them have their extra tasks marked "didn't need" so they aren't causing anyone to do unnecessary work.
I know exactly what you are talking about. I sometimes see a host of mine that has stopped crunching for some reason - a crash or BOINC has quit - and if this goes unnoticed for a few days, there can be deadline problems when the host is eventually restarted. It's very tempting to try to 'manage tasks through' rather than just aborting the ones at risk. There is always quite a percentage of tasks that don't get returned by others so you might well be aborting a task that could have been truly needed anyway. I tend to have a quick glance at the quorum partners to see if they are still returning regularly. If not, I don't abort my copy.
Quote:
One of the ones I aborted still hasn't been sent to anyone else; if I hadn't aborted it, I would have finished it by now. Oh well. Maybe Boinc needs an "ask server for deadline extension" button.
Funny you should mention that! :-). There is such a 'button' of sorts but it's not visible in the Manager anywhere. You need to 'talk' nicely to the server and it will grant you an extension if the deadline hasn't already been exceeded. Einstein has a BOINC server feature enabled called 'resend lost tasks'. All you need to do is inform the scheduler that the tasks you should be aborting have become 'lost'. You will be sent fresh copies with a new deadline. To do this you have to be comfortable with editing the state file (client_state.xml) and you need to fully understand how to do it correctly. So it's certainly not for the average volunteer.
One of the ones I aborted still hasn't been sent to anyone else; if I hadn't aborted it, I would have finished it by now. Oh well. Maybe Boinc needs an "ask server for deadline extension" button.
Funny you should mention that! :-). There is such a 'button' of sorts but it's not visible in the Manager anywhere. You need to 'talk' nicely to the server and it will grant you an extension if the deadline hasn't already been exceeded. Einstein has a BOINC server feature enabled called 'resend lost tasks'. All you need to do is inform the scheduler that the tasks you should be aborting have become 'lost'. You will be sent fresh copies with a new deadline. To do this you have to be comfortable with editing the state file (client_state.xml) and you need to fully understand how to do it correctly. So it's certainly not for the average volunteer.
Interesting, but I don't think I'll take the time to learn how to do that.
Anyway.... Yesterday, I finally updated the nvidia driver. I also took the time to go through all the available Windows updates and see what they were for, deciding to take some and reject others. Then I gave up on waiting for it to restart when it spent over an hour at 13% configured. When I checked it this morning, it had finished restarting, but most of the updates had failed to install! AAARGH!
I haven't seen anything in the pushed Windows updates about directx, but I plan to check for updates for that too.
After that, I'll just have to wait and see if dxgkrnl causes another crash.
David
Miserable old git
Patiently waiting for the asteroid with my name on it.
Okay, everything is done. The
)
Okay, everything is done. The ones that were returned late have validated; some of them have their extra tasks marked "didn't need" so they aren't causing anyone to do unnecessary work.
One of the ones I aborted still hasn't been sent to anyone else; if I hadn't aborted it, I would have finished it by now. Oh well. Maybe Boinc needs an "ask server for deadline extension" button.
Anyway, I will check into updating my video driver after work today. But since the span from the last bsod to this one was nearly four months, it may be a while before I know if the update helped.
David
Miserable old git
Patiently waiting for the asteroid with my name on it.
RE: Okay, everything is
)
I know exactly what you are talking about. I sometimes see a host of mine that has stopped crunching for some reason - a crash or BOINC has quit - and if this goes unnoticed for a few days, there can be deadline problems when the host is eventually restarted. It's very tempting to try to 'manage tasks through' rather than just aborting the ones at risk. There is always quite a percentage of tasks that don't get returned by others so you might well be aborting a task that could have been truly needed anyway. I tend to have a quick glance at the quorum partners to see if they are still returning regularly. If not, I don't abort my copy.
Funny you should mention that! :-). There is such a 'button' of sorts but it's not visible in the Manager anywhere. You need to 'talk' nicely to the server and it will grant you an extension if the deadline hasn't already been exceeded. Einstein has a BOINC server feature enabled called 'resend lost tasks'. All you need to do is inform the scheduler that the tasks you should be aborting have become 'lost'. You will be sent fresh copies with a new deadline. To do this you have to be comfortable with editing the state file (client_state.xml) and you need to fully understand how to do it correctly. So it's certainly not for the average volunteer.
Cheers,
Gary.
RE: RE: One of the ones I
)
Interesting, but I don't think I'll take the time to learn how to do that.
Anyway.... Yesterday, I finally updated the nvidia driver. I also took the time to go through all the available Windows updates and see what they were for, deciding to take some and reject others. Then I gave up on waiting for it to restart when it spent over an hour at 13% configured. When I checked it this morning, it had finished restarting, but most of the updates had failed to install! AAARGH!
I haven't seen anything in the pushed Windows updates about directx, but I plan to check for updates for that too.
After that, I'll just have to wait and see if dxgkrnl causes another crash.
David
Miserable old git
Patiently waiting for the asteroid with my name on it.