Benchmarking submethods

[jrichview]

I have a chunk of legacy code that queries data from SQL and then loops for each row calling 5 other methods some of which also do queries and populate some related data.

I'd like to benchmark each method in order to drill down on the most time-consuming code, but I assume I can't just add [Benchmark] attributes to each method. The library expects those attributes only on methods that it should call directly, correct?

But it would be quite difficult and time-consuming to set those sub-methods to be runnable by themselves the way they would run normally.

Or, would it make more sense to use the admittedly less accurate Stopwatch for a scenario like this?

[timcassell]

This sounds like a case for a profiler. If you want to use BDN for it, you will have to setup each method like you expect (#2484).