Message boards :
Number crunching :
validation algorithm
Message board moderation
Author | Message |
---|---|
PhonAcq Send message Joined: 14 Apr 01 Posts: 1656 Credit: 30,658,217 RAC: 1 |
When I look at my work queue, I typically have a few results waiting for validation. Today there were 5. If I then look at the server status, it typically has roughly the same number or zero. (Just now it was also 5.) I would think that I am a 'typical' user and so every active member should have a similar number waiting for validation, and that the number reported on the server status page would be approximately the sum of everyone, accounting for the inaccuracies in querying the database. So my quesiton is why is it that the number waiting for validation isn't a very large one? May this Farce be with You |
Keck_Komputers Send message Joined: 4 Jul 99 Posts: 1575 Credit: 4,152,111 RAC: 1 |
A workunit is not waiting for validation by the server standards until there are enough sucessful returns to possibly form a qurom (3+ at seti). The server status page is also only an approximation in most cases. Getting and keeping accurate numbers would incur to great a performance penalty. BOINC WIKI BOINCing since 2002/12/8 |
Ingleside Send message Joined: 4 Feb 03 Posts: 1546 Credit: 15,832,022 RAC: 13 |
Only when atleast 3 "success"-results for a wu have been reported is the need_validate-flag set by the Transitioner. The status-page counts how many "need_validate" there is, meaning how many is waiting in the Validator's queue to be tried validated. Showing this on status-page will reveal if there is a bottleneck in validation or not. It is not a count of all results "pending", since most of the time they're waiting on 1 or 2 other results before can be tried validated. While it could be interesting to get a detailed count of how many is pending waiting on 1 or 2 results, or for that matter got 3 but failed and waiting on a 4th, none of these info will say anything about the servers having a bottleneck or not. Actually, if something should be added to the status-page, my 1st choise would be a count of how many "success"-results reported last 24 hours, maybe with a count of how many reported as errors, past deadline, failed validation and "no consensus yet". |
PhonAcq Send message Joined: 14 Apr 01 Posts: 1656 Credit: 30,658,217 RAC: 1 |
I just checked the credit on one result http://setiweb.ssl.berkeley.edu/workunit.php?wuid=15778357 It shows two completed calculations with a 'normal' amount of time of several thousand seconds, another result with 'zero' time (20 secs), and a fourth calculation still pending, with no results returned yet. Nevertheless, two of us got credit. I thought you needed three results to get credit. I have to think the 20 second result is bogus, so only two results could have agreed at this point. Is there a simple explanation? May this Farce be with You |
©2024 University of California
SETI@home and Astropulse are funded by grants from the National Science Foundation, NASA, and donations from SETI@home volunteers. AstroPulse is funded in part by the NSF through grant AST-0307956.