Here's another rather simple test case that exhibits an unusually long calculation time. In this case it is not FullSimplify, but the newly introduced Div in spherical coordinates. Since one cannot compare its calculation time to M8 (where it simply doesn't exist), I compared it to evaluating the divergence step by step using the well-known formula. The step by step evaluation runs quickly in M9, including simplification (it is 0 in the end, which is of course correct), but Div appears to hang at full CPU usage. (Actually I started the calculation before writing this post and it's still calculating.)
Anyone to test and confirm this?
Attachment: div.nb, URL: ,