They would have to recompute the chance of a new dotted red line after every tick of the green line, to see whether the dotted red line would have a new higher energy and move or not. If so, it would have to then redraw the graph. All of that constant checking and regraphing the current plot would be a huge waste of crunching time.

You also don't know what would happen to the dotted red line on the right. Horizontal is assumed based on the data to the left but there may have been previous lower energy on the right. The red line would have to be partially new and partially old. A lower green energy could always be found, meaning another dotted red line recompute to that new point.

I think it only checks after the green line has finished, a checkpoint has been saved and it is ready to start over.

That question I can answer. All the scaling, etc, is based on your work unit alone. Yes, that means a lot of poor or so-so possibilities get returned to the server, but it can power through all the possibilities quickly enough. This way if an error causes an impossibly good answer for one work unit, we do not lose all the realistic possibilities in other work units.
mycrofth

I'll try to clear a few things up. My definitions aren't going to be completely accurate, but you should get an idea of what's going on. I'll let the scientists give more detailed explainations of the terms.

* The graphs are for the current workunit only.

* The length of the graph represents a single, complete docking. A docking is composed of thousands of docking attempts over a "population".

* A "population" is composed of (typically) hundreds of "individuals". A single docking calculation iterating over the population represents a generation. The best (lowest) energy of the generation is plotted as a dot on the green graph.

* The dotted red line represents the graph of the best docking run seen so far for this workunit. "Best docking" meaning the docking that had the best (lowest) energy.

* The green line represents the current docking. It may or may not dip below the red line. At the end of a docking, if the energy of the green line is less than the red line, the green graph becomes the new red graph.

As for not seeing anything plotted, that's currently possible because we only display dockings that fall below a certain threshold. As long as the percentage keeps climbing, the workunit is running fine. I know it's a bit confusing, so I have changed this behavior and you should now always see something plotted in the C panel. This change will be deployed very soon.

I hope I didn't confuse folks any more.

I had been using version 4.0.3.1 and will continue to finish those packets as they come up. Today's new program download is on the order of 915,000 bytes and contains a new version 4.0.3.2. We can watch for the new threshholdless behavio(u)r in the Best Docking Energy graph.