We don't have a solution to the validation problem. It is really annoying for volunteers and for us, but we could not figure out why identical output files fail to validate. The good news is that this only affects the present app, and since we do not expect to release any new WUs with this app, this problem should simply become irrelevant as soon as a new app is deployed. The new app will use a completely different validation approach (SQLite database files instead of plain ASCII data files). If the problem is still present in the new app, we will get back to it and do everything to solve it, you can be sure about it.
As soon as we start testing the new app, all the unfinished WUs will be deleted, and no credit will be granted for those WUs. Don't take this as a punishment, it's just that assigning a given credit would require serious BOINC database manipulations (select users with non-validated WUs; make sure this is actually related to the validation problem; decide how much credit assign; assign it), and this is a big chance to introduce inconsistencies in the database. For us, the database and all the php files accessing and modifying it are a black box, so we try to keep manual interventions to a bare minimum.
Now some good news: the new app is looking good, even if still not ready for deployment (maybe 3 months to go, mainly due to other issues now requiring our attention). It will take over the old app, and run on real data (real sky coverage by real NEA surveys, with new data coming in every night). WUs should run much faster on this app, and we will try to keep average crunch time on the order of hours on most computers.
Pasquale
Discussion Zen 5 Speculation (EPYC Turin and Strix Point/Granite Ridge - Ryzen 9000)
Discussion Intel current and future Lakes & Rapids thread
Facebook
Twitter