Questions and Answers :
Windows :
First invalid tasks, but wingmen results match
Message board moderation
Author | Message |
---|---|
JakeTheDog Send message Joined: 3 Nov 13 Posts: 153 Credit: 2,585,912 RAC: 0 |
Hello, I don't always check my results daily, so I may have not seen some before, but today I have finally seen my first invalid results. However, when I look at the wingman in 1 work unit, he/she has identical results and is inconclusive. In the other work unit, I have 2 wingmen both with identical results, but 1 is invalid while the other is inconclusive. Both WU's have another person with "in progress." Links below: http://setiathome.berkeley.edu/workunit.php?wuid=1669857527 http://setiathome.berkeley.edu/workunit.php?wuid=1669873326 Seems very strange. 1) So is there something in the results that explain the validation status, or is there something wrong with the status? 2) What % of invalid results is "normal" for a stable PC? |
OzzFan Send message Joined: 9 Apr 02 Posts: 15691 Credit: 84,761,841 RAC: 28 |
There is more to the results than what you see in the STDERR output file visible on the website (e.g. Spike, Pulse, Triplet, Gaussian, Autocorr, etc.). If the results don't match strongly, a validation inconclusive will occur. If the results don't match at all, or there is an error in processing, an invalid will occur. As to what percentage of invalid or inconclusive is acceptable, that really depends on a lot of factors, but I'd say about 0-1%. If you're getting a lot of errors, I'd check to make sure you're not overheating (the excess heat can cause computational errors), or if you're overclocking, make sure you're not over-volting too much. |
Jeff Buck Send message Joined: 11 Feb 00 Posts: 1441 Credit: 148,764,870 RAC: 0 |
It's not you or your machine. Those tasks ran in a "Legacy setiathome_enhanced V6 mode" and therefore didn't produce an Autocorr result. See the Strange, utterly strange MB v7 run as MB v6 with MB v7 app.... thread over in Number Crunching for additional info on the problem. |
JakeTheDog Send message Joined: 3 Nov 13 Posts: 153 Credit: 2,585,912 RAC: 0 |
Yeah, that is weird that there is no autocorrect #. That first WU now has 5 invalids. |
Wisesooth Send message Joined: 1 Oct 14 Posts: 6 Credit: 3,138,923 RAC: 0 |
I strongly suspect that Microsoft may be (at least part of all) the cause of the problem. If its auto-update service creates a restart at the end of an OS update, it does not seem to play very nicely with BOINC. If BOINC projects leave cache in RAM while suspended, or do not have checkpointing to recover from open-file restarts, a Microsoft restart may corrupt the project file. Automatic Java updates may also cause this to happen. Not sure this is going on, but I smell a rat in the woodpile. |
©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.