climateprediction.net home page
Bug in CPU time?

Bug in CPU time?

Questions and Answers : Getting started : Bug in CPU time?
Message board moderation

To post messages, you must log in.

AuthorMessage
old_user31627

Send message
Joined: 28 Nov 04
Posts: 5
Credit: 352,070
RAC: 0
Message 9397 - Posted: 16 Feb 2005, 15:11:31 UTC

I had a unit that finished and deleted total time in "Summary of workunit". I think there is a bug there.

Last Trickle Information for Result # 409412 is:
13 Feb 2005 11:13:54 86573 409412 2245_100117030_2 3 259248 2526436 3.2484
13 Feb 2005 11:13:54 86573 409412 2245_100117030_2 3 248446 2490948 3.2479

But the information of the result 409412 is
CPU time: 0.00
Received: 14 Jan 2005 20:06:15 UTC
Server state: Over
Outcome: Unknown
Granted Credit: 6805.26
ID: 9397 · Report as offensive     Reply Quote
Profile Thyme Lawn
Volunteer moderator

Send message
Joined: 5 Aug 04
Posts: 1283
Credit: 15,824,334
RAC: 0
Message 9400 - Posted: 16 Feb 2005, 16:11:18 UTC

Although your last trickle was sent on 13 Feb but the result was received by the server on 14 Jan. Did you restore the model from a backup after a crash?

BOINC prevents an entry in the result table from being updated once its received time has been set. This doesn't prevent a CPDN model from continuing to completion, accumulating more credits and uploading the result files (as yours has), but the table doesn't get updated to reflect the new outcome. Having said that, I know that result handling is non-standard in CPDN, but this part may not have been changed.

If it hasn't been changed, here's a suggestion for the project team to prevent confusion.

If you receive a new trickle for a result that has already been received would it be possible to set the server state back to 'in progress' and the result to not received?
"The ultimate test of a moral society is the kind of world that it leaves to its children." - Dietrich Bonhoeffer
ID: 9400 · Report as offensive     Reply Quote
old_user31627

Send message
Joined: 28 Nov 04
Posts: 5
Credit: 352,070
RAC: 0
Message 9418 - Posted: 16 Feb 2005, 18:43:58 UTC - in response to Message 9400.  

> Although your last trickle was sent on 13 Feb but the result was received by
> the server on 14 Jan. Did you restore the model from a backup after a crash?
No. I didn't do anything manually.

Yes I think something have to be done.
First. CPU time was updating well until the end.
Second."Summary of workunit" should update somehow or stop the unit somehow.

How do project team get to know of this kind of bugs? Is there a place to report bugs?
ID: 9418 · Report as offensive     Reply Quote

Questions and Answers : Getting started : Bug in CPU time?

©2024 cpdn.org