Yes I think you are right it would make more sense for it to have been a decoupling cap. I have access to a rework station & would have attempted to stick it back on should the chip have failed to post. I've tested everything everyway I can think of & everything looks to be working as it did before & better thanks to running cooler so looks like I got away with it.
4800 Test
http://i106.photobucket.com/albums/m...0oCAmbient.png