Are all the 'invalid' results still usefull or does the file itself become corrupted? Zero credit would not be my first concern at this point but I do care if a (large) part of this new project turns out to be a waste of energy before the problem is solved.

Greetings everyone,

Thank you for the initial report and detailed ways to recreate it. I have implemented a fix on our end that will hopefully resolve this issue and get more results validated.

The basic problem was that during intermediate uploads, the files were not compressed when sent back to our server. Which is how we planned for them because the larger of the two files is binary and doesn't compress much. However, when BOINC completes a work unit, it compresses the intermediate upload flies and they were not comparing matching part of the trickle messages coming back from members. The fix is server side, and expect to see more results turning valid from this.

Thanks,
-Uplinger

Unfortunately with the way things are on the system, I can not go back and validate older work units, this would disrupt generation of work units. This is a change going forward. Members should see a lot fewer invalids because of the changes to the trickle message handler.

Thanks,
-Uplinger

Never come across this before in my 8+ years at WCG, no other BOINC projects do this, do they ?

I now have 21 Invalids , a waste of science and a waste of power and money - I don't want the points but the hours towards a badge is useful .. and the chances of that are, oh well ...

Interesting in that it didn't initially show any credit on the invalids but I got up this morning to find credit granted. What causes the actual delay in credit being granted?