In the last target a library of about 2.85 million compounds were put against the target docking simulations. If same, then nearly 9 million are needed to be run. Since the work is chopped up so it does not run too long for the average cruncher i.e. when the tasks do take too long the compound dockings are split up over multiple WU's, there thus no telling how many are needed.

You could follow the current task name number in your client and then derive from there how far we are. My Linux quad is now working on 00035270_n1_0000_0. The last part indicates if there are more for 1 compound, but probably not as the mean run time is just 3.37 hours. At any rate the techs test a set of a dozen or so for each new batch [through the production pipe] to find out how long they run and then decide on sizing them up. These test jobs can be recognized by an anomalous prefix to the WU name, so the techs can find them back easier in the database.

kind of.

PS, this is a sample name of aWU of the previous target towards the end:

HFCC_s2_02844035_s2_0001

Sek,

With the setup of the initial work units and how many loops are allowed in Autodock (without doing a recompile) we split all the work units in half. Unless the work units become more difficult, it is probably a safe assumption that the 2.85 million per target * 2 for the split is on the low side. Then multiply by 1.25 (to account for possible resends and host checking) to get the high side of estimated results sent out.

This is of course estimations as we still have not tested the other two targets yet.

-Uplinger

dkt, hopefully it will help you to get a blue badge for HFCC
I've changed my WCG preferences to crunch about 40% HFCC - hope to reach 8years before it's all over.
Since yesterday I'm about 4days closer

Make that two of us who want to know....

I have a feeling these three targets will go much faster than expected. Plenty of people badge hunting, badge padding for the 10 year, and a very worthwhile project.

Happy to be crunching it again.