Message boards :
AstroPulse :
Invalid, little time, lots of credit
Message board moderation
Author | Message |
---|---|
Send message Joined: 11 May 13 Posts: 36 Credit: 1,001,497 RAC: 0 ![]() |
I just came across a computer that consistently produces invalid results after short run times, none the less it keeps getting lots of credit... Is there a problem with the validator and/or with the credit mechanism? |
Send message Joined: 29 May 06 Posts: 1037 Credit: 8,440,339 RAC: 0 ![]() |
Link to the host? Claggy |
Send message Joined: 29 May 06 Posts: 1037 Credit: 8,440,339 RAC: 0 ![]() |
The host in question has a Intel HD2500 GPU and is running OpenCL 1.1 drivers, while it doesn't produce the correct result, Since this is a Beta project it is important to find out what works and what doesn't: host 69098 I've just reported the misreporting of the OpenCL driver version to the boinc_dev list (Where the server displays 1.01, it should say 1.1) Claggy |
![]() Send message Joined: 14 Feb 13 Posts: 606 Credit: 588,843 RAC: 0 |
I just came across a computer that consistently produces invalid results after short run times, none the less it keeps getting lots of credit... No, since this is beta, there is a high proportion of tasks that fail because the app is still being developed. Eric periodically runs a credit granting script, which assigns credit even to invalids. reasoning, it's not your fault, your host did work for the project, you shouldget credit for that, even if the result didn;t fit. Some rigs that are actually misbehaving and need looking at are a small price for keeping the majority sweet. NB the script only gets run when there is reason to believe that problems were due to an app. Since the Android app is still causing lots of headaches... The script does run globally though and can't distingusish between runaway rigs and runaway apps. Beta testers should actually be monitoring their tasks more closely, especially when involved in active development and pipe up with any problems, but alas only very few do. Of course also a rig that is quietly ticking away is of help, but people should occasionally check that everything is running smoothly. Sorry, preaching to the choir :D IOW the validator is fine and Credit is CreditNew with all the problems incurred to the amount of credit assigned. Actually, I'd have thought the invalids would be granted the same credit as the others - that looks more like a (random) fraction. Interesting. Unfortunately Eric is incredibly busy and deosn;t really have the time to look into sich things. I suspect it's just CreditNew with extra screw factor due to belated credit granting, but I'll try and ask Eric what the script looks like, maybe there really is a problem. Truth be told, we aren't really fussed about credit here, so I doubt Eric will make time for that. More about keeping people happy by giving credit a bit more generously. A person who won't read has no advantage over one who can't read. (Mark Twain) |
©2021 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.