climateprediction.net home page
Sudden increase in time to complete

Sudden increase in time to complete

Message boards : Number crunching : Sudden increase in time to complete
Message board moderation

To post messages, you must log in.

AuthorMessage
old_user60144

Send message
Joined: 4 Mar 05
Posts: 6
Credit: 65,136
RAC: 0
Message 16686 - Posted: 20 Oct 2005, 11:38:48 UTC

I just changed to Boinc 5.2.1 a few days ago and 5.2.2 this morning and all went well. My projects, including a sulphur model, are going right along.

But... prior to changeover my percent complete was 48 with 735 hrs done and 779 still to go. Immediately after the change the time to completion shot up to 1743! Updating to 5.2.2 didn\'t change anything and I am now at 812 hrs done, 53.2% complete, and 1646 still to go.

Does anyone have any thoughts on what is going on?

Thanks, Karl N
ID: 16686 · Report as offensive     Reply Quote
old_user272

Send message
Joined: 6 Aug 04
Posts: 58
Credit: 1,286,603
RAC: 0
Message 16687 - Posted: 20 Oct 2005, 12:34:16 UTC - in response to Message 16686.  


Does anyone have any thoughts on what is going on?


The 5.2.x versions calculate the end time in a different way, one that takes into account how long previous results took. This works well for \"normal\" projects but not for projects like CPDN with it\'s long duration work units.

Ian

ID: 16687 · Report as offensive     Reply Quote
old_user60144

Send message
Joined: 4 Mar 05
Posts: 6
Credit: 65,136
RAC: 0
Message 16688 - Posted: 20 Oct 2005, 13:11:14 UTC - in response to Message 16687.  


Thanks Ian,
May I then be confident that the percent complete value will let me estimate the final completion date reasonably accurately?
Karl
ID: 16688 · Report as offensive     Reply Quote
Les Bayliss
Volunteer moderator

Send message
Joined: 5 Sep 04
Posts: 7629
Credit: 24,240,330
RAC: 0
Message 16689 - Posted: 20 Oct 2005, 13:41:53 UTC

For a sulphur model, record how long between trickles, then multiply that by 120 for the full model. For a slab model, multiply by 72.

ID: 16689 · Report as offensive     Reply Quote
old_user272

Send message
Joined: 6 Aug 04
Posts: 58
Credit: 1,286,603
RAC: 0
Message 16690 - Posted: 20 Oct 2005, 13:44:20 UTC - in response to Message 16688.  


May I then be confident that the percent complete value will let me estimate the final completion date reasonably accurately?


It should be in the right area. You can also calculate it from the timestep count and timesteps per second displayed in the visualisation.

Ian

ID: 16690 · Report as offensive     Reply Quote

Message boards : Number crunching : Sudden increase in time to complete

©2024 cpdn.org