The above code is optimized for diagram clarity, but it executes unneeded calculations. If we place the "rare" extra calcuations inside a case structure, we get a bit faster (since we can assume that a jump is a relatively rare event!).
The following code is entirely competitive with the stock "unwrap phase".

With debugging disabled and inlined, it's actually ~5% faster than stock (tested on a 2M sine wave as in the previous picture), probably because we don't have the subVI calling overhead. For all practical purpose, they are the same. (further very small improvements are possible, but probably not worth it).
Message Edited by altenbach on 06-20-2009 11:31 AM