I *may* have figured out the problem. After a bunch of reading online, browsing various forums, and google searching I started looking at how the pawl contacts and turns the cylinder. I think it's a question of mechanical (dis)advantage in combination with a cylinder that isn't turning as freely on the shaft as it should be. Never having owned a revolver before I wasn't sure just how smoothly/easily the cylinder should be turning. It occurred to me to try to push the cylinder in the same manner as the pawl. Doing that I could tell that the cylinder became very hard to turn by pushing on the bearing surfaces of the star gear. It's not easy on any of the six positions but at least half of them it's essentially impossible to turn the cylinder by "hand" in this manner.
Short story is, I stripped it down to where I removed the crane assembly, even found out how to remove the frame latch and ejector rod, but then found that on this GP100 that's as far as I go without a specialized tool to completely remove the cylinder from the shaft. I was able to test my theory by applying a small amount of gun butter to the point where the cylinder meets the shaft and letting that soak for a while. The cylinder is much smoother and freer now and there's no more "freeze up" on three cylinders. There's still a definite stiffness on three so I think I'll still send it back to Ruger along with a note explaining my theory. I suppose I could go the route of soaking the crane and cylinder to fully clean it, but I'd just as soon have Ruger check it out throughly. Maybe the bearings or shaft are in need of work.
Long post, but figured it might be interesting to some. I'll post again once I get the gun back from Ruger.