Duration Correction Factor

log in

Advanced search

Message boards : Science : Duration Correction Factor

Author Message
Profile Dirk Sadowski
 
Send message
Joined: 28 Mar 10
Posts: 21
Credit: 217,666
RAC: 41
Total hours: 620,552
Message 706 - Posted: 28 Jun 2012, 14:05:34 UTC

Since a few days my machine calculate WUProp@Home work units...

The estimated time of the work units was correct a long time (~ 3 hours).

Then I made a reboot of my machine and then the estimated times are too high (~ 3 days).

I didn't looked to the Duration Correction Factor before, but now it's ~ 28.
Normally it should be ~ 1.

Is this a known problem?

____________

Link
 
Avatar
Send message
Joined: 20 Jun 12
Posts: 122
Credit: 334,892
RAC: 113
Total hours: 187,006
Message 707 - Posted: 28 Jun 2012, 19:22:19 UTC - in response to Message 706.

Then I made a reboot of my machine and then the estimated times are too high (~ 3 days).

You mean, after 10 WUs were validated and the "server side DCF" kicked in? Yeah, had that on all my machines. Not nice, but with 1 task/machine limit it actually doesn't matter.



I didn't looked to the Duration Correction Factor before, but now it's ~ 28.
Normally it should be ~ 1.

You have a DCF for WUProp@Home? On all my machines the entire scheduling part (to which the DCF belongs) of project's properties is missing... all I have is "General" and "Credit".
____________
.

Link
 
Avatar
Send message
Joined: 20 Jun 12
Posts: 122
Credit: 334,892
RAC: 113
Total hours: 187,006
Message 708 - Posted: 29 Jun 2012, 12:59:24 UTC

OK, there is a DCF in the client_state.xml. It's adjusting itself slowly to the new estimates send by the server. The first WU had IIRC an estimated runtime of just few minutes, so that's where the high DCF comes from.
____________
.


Post to thread

Message boards : Science : Duration Correction Factor


Home | My Account | Message Boards | Results


Copyright © 2024 Sebastien