So, I just got into the lab this morning and go figure...the machine, after days of zippo communication, with absolutely no changes or tweaks on my part, has suddenly "been allowed" to upload everything in the massive upload queue and is now pulling down new WUs. So much for the idea that "BOINC will not request work if the number of uploads is greater than 2 * number of processors."

There is nothing to test here when the variables on the S@H side are constantly changing. And the users are kept completely in the dark as to what really happens during these weekly *cough* maintenance blackouts.

I give up. Since we (the power crunchers) appear to be absolutely powerless to do anything proactive to help things along, I make this final recommendation.

Create new and separate code sets for "Johnny average screensaver guy" and the power crunchers. The way it stands, the code seems to be trying to do too much.

You know, I used to be really stoked back in the 1999-2000 timeframe when I had a DOS version of SETI classic I could run in a batch file, one instance for each core or logical proc. It was a piece of cake and you always knew exactly what was going on or at least could isolate things much better.