Michael,

Thanks for the reply. I can see that testing is the only sure-fire method to compare the performance, but there must be some indication that the switch from memory to tempdb has been made. I've inherited support for this application, notorious for its poor performance, and I'm trying to get my arms around all the targets of opportunity to determine where to strike next. That is why I want to determine if the switch to tempdb is happening before dedicating time that I can use in other tuning methods.