Looked thru the checkpoint log entries and count about 32 attempts per job. I think we need the techs or watowich to tell us how much is packed. WCG nor a scientist would engage in an open ended project, thus for now either the 1 year calendar or 50,000 CPU years does feel like a realer number. That said, the estimates have been off pretty much all the time (speak running considerable longer). Look at the HCMD story e.g.

Thanks for coming to rescue me esoteric17.

I have no doubt that watowich is not trying to confuse us but there are obviously some parameters that we are missing, and that is what I asked for.

Regarding the 50,000 CPU years figure I would not give it much importance when trying to find how long it will take us to complete this project (in real human calendar months): I consider that such preliminary figures can be used only for guessing if the project is small, big or huge, nothing more precise.

And, of course, I would not try to extrapolate the current "ramping-up" daily figures either: most people don't know yet that there is a new project that they can select. And only if/when they have migrated to Boinc. Not forgetting either the Boinc agents which run with many days of queued WUs and are only just about to start DDDt eventually.

Thank you too, ssnegi, for explaining how it may be possible to reduce the raw theoretical number of cases to a more realistic number of real WUs.

Cheers. Jean.

We've also used (part of? "fragment"?) ZINC for FAAH:

http://fightaidsathome.scripps.edu/

Experiment 7a2: 93% Completed:
x2BPW vs ZINC Fragment Library (exo site)

Experiment 7b1: 93% Completed:
x2AZ8 vs ZINC Fragment Library (active site)

Experiment 7b2: 93% Completed:
x2AZ8 vs ZINC Fragment Library (exo site)

agree.

Between 15 and 20 mn most of the time for me (P4 2.4 GHz).
With a few WUs between 15 and 10 mn.

Cheers. Jean.