I work for Sun Microsystems, and it's not that rare. Perhaps it's because I see the errors when they happen on all the systems we have across North America as people report them, but I get at least a few hits a day regarding single bit ecache errors. I would say they happen probably once, in a five year period, for each processor. Well, maybe less. Actually, the more I think about the amount of systems we have out there, and that there's a few a day... Hell, I guess it is pretty rare. But it happens!