Inconsistency in shortfall calculation?


Advanced search

Message boards : BOINC client : Inconsistency in shortfall calculation?
Message board moderation

To post messages, you must log in.

AuthorMessage
Erik

Send message
Joined: 18 Oct 07
Posts: 2
Message 13125 - Posted: 18 Oct 2007, 14:45:52 UTC
Last modified: 18 Oct 2007, 14:47:55 UTC

Looking at the rr_simulation debug output, I noticed something that looks inconsistent. It's in the way shortfall is calculated for projects with no existing tasks, vs those that have tasks but finish within the work buffer time. Basically it looks like the former have their shortfall modified by ocf (on computer fraction?) while the latter don't.

Is that intended or is it a bug (or am I misunderstanding something)?

This is in version 5.10.20 on WinXP.
ID: 13125 · Report as offensive
Erik

Send message
Joined: 18 Oct 07
Posts: 2
Message 13668 - Posted: 8 Nov 2007, 16:03:30 UTC

Hi,

No replies on this thread yet... I suspect it is a bug; not a particularly important one, but probably trivial to fix.

Example: I just saw LHC@Home go from calculated shortfall of 62ksec just before finishing its last result, to a calculated shortfall of 38ksec after. That corresponds nicely to the ~0.6 ocf for this machine.

See above for details. Now running 5.10.28 on WinXP.
ID: 13668 · Report as offensive

Message boards : BOINC client : Inconsistency in shortfall calculation?

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.