@ExtraTerrestrial Apes

The solution would be to have 2 different hostname for a single device.
That way you could have 2 different profile in same computer, one with OPN@GPU only and another for CPU only projects.
Of course you may have 2 different Boinc's data directory too.

There is another way, one I use myself. Set one profile e.g. school, to Open Pandemics GPU only WU's (no use of CPU, but GPU allowed), then set another profile e.g. work, to "everything else you want to crunch on CPU".

On the PC concerned set WU cache size to, say, 2 days. Now go do Device Manager in WCG website and apply first the "Work" profile to that PC. Click the update button in the BOINC client on the PC until it is full of CPU WU's. Then swap the device profile to "School" in the WCG device manager.

Now you have a full cache of CPU WU's to process, but the machine will only ask for OPN1 GPU WU's from now on.

Every 2 days or so swap profile back to worjk for about 10 mins to allow it to fill up again with CPU units, then swap back to school. This has the beneficial side effect that on a multi thread CPU there are regular cpu units completing and triggering a request for new work. Which will try to get GPU units.

Great question and a keen eye! Most of our packages, in addition to the candidate ligand and a static protein target, also feature a flexible part of the protein, which, from the simulation's perspective, is essentially a second ligand. This is the second set of atoms you can see in the screensaver graphics.

The so called flexible residue can help to more realistically simulate ligand behavior by allowing better steric fit and for some of our packages is even used to model chemical reactions between the ligand and protein.

WCG is not restarted yet, the WUs presently available are only for testing the system,
If you are worried about getting credit I would suggest running a different project until WCG is fully up and running again.

Thanks for the info. I was wondering what was going on with the WCG points system.