My understanding of distributed computing is that we are substituting volume for speed -- that is, a "supercomputer" like IBM's "Blue Gene/L" might do a job in one day, but is a million time faster than my PC, so we break the job into a million pieces and get a million people to each contribute one day of computing to get the job done in the same time. Or, we get 1,000 people to donate 100 days of computing. I know thats a bit oversimplified.

My question is this: is there any estimate of how long it would take a machine like Blue Gene, or Cray 2 or Deep Blue or these other supercomputers to do the Fighting Aids or protein folding projects?

Perhaps, in addition to volunteering our computers, we might also lobby our government to build a couple of Blue Gene-type machines dedicated solely to medical research. My understanding is that the existing machines, one at Lawrence Livermoore I believe, are most often used for climate modeling (fair enough -- a good cause), and nuclear weapons research (well, I'll leave that to your won evaluation).