Message boards :
Number crunching :
Data losses in the "statistics" or simple "WU cancer" ?
Message board moderation
Author | Message |
---|---|
Ramon Kemnitzer Send message Joined: 9 Feb 00 Posts: 3 Credit: 2,720,953 RAC: 0 |
I have registered today a problem with the counting of the workunits in the "Statistics": Computer-ID 1058955 05.08.2005 - 283 units 08.08.2005 - 239 units (UTC 10:00 a.m.) 08.08.2005 - 197 units (UTC 04:30 p.m.) 08.08.2005 - 192 units (UTC 08:00 p.m.) In average the machine needs 2 hours 30 minutes per WU. It must become more !!! I have seen it for some more of my machines, mostly AMD. Anybody elsewhere seen the same phenomenon ? |
Ramon Kemnitzer Send message Joined: 9 Feb 00 Posts: 3 Credit: 2,720,953 RAC: 0 |
I have registered today a problem with the counting of the workunits in the "Statistics": Now I found some other computers of my farm (the same dates) figures in the field "results" 1189236 197 222 201 191 1058936 122 102 84 80 1075379 85 88 78 81 1081431 77 81 66 68 What's going on? Per aspera ad astra. |
Charlie1 Send message Joined: 2 Aug 00 Posts: 28 Credit: 15,684,838 RAC: 0 |
If I'm right, the stats only shows the units that are still active in the system. Once they are validated and credit is granted the results are stored and they are removed from the system to make room for new units. So the number of units that shows in the list will vary but your credit should be fine. If this is not correct I'm sure someone will point it out. :) |
Ramon Kemnitzer Send message Joined: 9 Feb 00 Posts: 3 Credit: 2,720,953 RAC: 0 |
If I'm right, the stats only shows the units that are still active in the system. Once they are validated and credit is granted the results are stored and they are removed from the system to make room for new units. So the number of units that shows in the list will vary but your credit should be fine. Hello Charlie1, Thanks for your posting. I miss all the units earlier than 17.07.2005. It shows in the "192" also validated units and granted credits. We will see what is the problem - I think the "Results" stands for a "Total". |
Paul D. Buck Send message Joined: 19 Jul 00 Posts: 3898 Credit: 1,158,042 RAC: 0 |
BOINC does not maintain a count of work units processed. It was felt that this was irrelevant when you have projects like LHC@Home where the work units are variable in size but not distinct in any other feature. Also, it is not cross project. The number of work units listed on each computer's summary page is only a count of how many are CURRENTLY Resident. As the db purge and file deleter operate, this number will go up and down. If you want to count, well, you have to use a third party tool like BOINC View to log the results processed and maintain a count. For an example I use the BOINC View logs and then update the database with the output files once a week. Note however, that the numbers on that page are not the absolute totals, but only the totals since I started capturing the logs ... |
©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.