climateprediction.net home page
Credit Issue Resolved. Thanks and Well Done!

Credit Issue Resolved. Thanks and Well Done!

Message boards : Number crunching : Credit Issue Resolved. Thanks and Well Done!
Message board moderation

To post messages, you must log in.

AuthorMessage
Profile Nick Perry
Avatar

Send message
Joined: 12 Jul 05
Posts: 11
Credit: 528,541
RAC: 0
Message 47122 - Posted: 19 Sep 2013, 11:15:45 UTC

My puny little machine seems to be behaving itself a lot better now.

Thanks to all for the effort to fix this irritating issue.
Prior to the debacle of the 19th century industrial revolution, Earth�s ecology had been an object of particular admiration for commentators in the Western Spiral Arm of the Galaxy .
ID: 47122 · Report as offensive     Reply Quote
metalius
Avatar

Send message
Joined: 28 Nov 06
Posts: 89
Credit: 11,328,674
RAC: 2,783
Message 47134 - Posted: 20 Sep 2013, 12:37:24 UTC

Credits are calculated, but only once.
So, it looks like - the very old bug, when credits are calculated only sporadic, still exists.
ID: 47134 · Report as offensive     Reply Quote
Richard Haselgrove

Send message
Joined: 1 Jan 07
Posts: 925
Credit: 34,100,818
RAC: 11,270
Message 47135 - Posted: 20 Sep 2013, 13:35:26 UTC - in response to Message 47134.  

Credits are calculated, but only once.
So, it looks like - the very old bug, when credits are calculated only sporadic, still exists.

I'm seeing a second update, roughly in line with my 'RAC' or daily average, making its way slowly through the very long credit calculation and reporting pipeline: it's reached "...from the daily update + 622.08 since then" in the BOINCstats system.

CPDN credit is a long and slow process. A huge script - the one we've been missing - has to work its way through the entire database: we worked out yesterday that it took at least five, possibly six, hours. Because of the problem with double-counting Beta stats (noticed after the first run), it had to be run twice on Wednesday - once on an overnight schedule, then again manually at lunchtime. After that has finished, the updated stats have to be exported in a format the external sites can use, and after that it's up to the external sites themselves to collect, process, and display the results.

None of these processes can be allowed to overlap, else one or more of the scripts will be working on partial - corrupt - data. Better that we miss a day and get it right, than allow that to happen.

Now that the admins have an idea how long the various steps take on the new hardware, they will be able to work back from the optimal 'collection' time, and schedule next the export script, and then firstly the update script, to start their respective tasks at the right time of day (or night). It may take a few attempts to get that timing absolutely right, but I think the admins are to be commended for the progress they've made this week.
ID: 47135 · Report as offensive     Reply Quote
Profile mo.v
Volunteer moderator
Avatar

Send message
Joined: 29 Sep 04
Posts: 2363
Credit: 14,611,758
RAC: 0
Message 47144 - Posted: 21 Sep 2013, 11:15:48 UTC

Metalius, it was decided a year or two ago (I can't remember exactly), long before the recent credits problem, only to run the credit script I think three or four times per week. That was a deliberate decision, not a bug. I don't know whether Jonathan and Andy will now want to run the script every day or just a few times each week. If they take the alternate-day decision it will be understandable because of the long-running nature of the script, as explained by Richard.




Cpdn news
ID: 47144 · Report as offensive     Reply Quote
metalius
Avatar

Send message
Joined: 28 Nov 06
Posts: 89
Credit: 11,328,674
RAC: 2,783
Message 47162 - Posted: 25 Sep 2013, 5:05:20 UTC - in response to Message 47144.  

Information accepted, thank You!
ID: 47162 · Report as offensive     Reply Quote

Message boards : Number crunching : Credit Issue Resolved. Thanks and Well Done!

©2024 climateprediction.net