Thread 'We've added two mechanisms that give better estimates of result completion time (and thus better scheduling and fewer missed deadlines): 1) we track the 'CPU efficiency': the average fraction of CPU time that BOINC apps receive while they'

Message boards : News : We've added two mechanisms that give better estimates of result completion time (and thus better scheduling and fewer missed deadlines): 1) we track the 'CPU efficiency': the average fraction of CPU time that BOINC apps receive while they
Message board moderation

To post messages, you must log in.

AuthorMessage
ProfileDavid Anderson
Volunteer moderator
Project administrator
Project developer
Avatar

Send message
Joined: 10 Sep 05
Posts: 726
Message 29976 - Posted: 24 Jun 2005, 0:00:00 UTC

We've added two mechanisms that give better estimates of result completion time (and thus better scheduling and fewer missed deadlines): 1) we track the 'CPU efficiency': the average fraction of CPU time that BOINC apps receive while they're running (this can be less than one on systems that run compute-intensive tasks other than BOINC); 2) we dynamically maintain a per-project upper bound on the ratio of actual CPU time to predicted CPU time.
ID: 29976 · Report as offensive

Message boards : News : We've added two mechanisms that give better estimates of result completion time (and thus better scheduling and fewer missed deadlines): 1) we track the 'CPU efficiency': the average fraction of CPU time that BOINC apps receive while they

Copyright © 2024 University of California.
Permission is granted to copy, distribute and/or modify this document under the terms of the GNU Free Documentation License, Version 1.2 or any later version published by the Free Software Foundation.