Good hands-on references for VS2010 with NSight 3.0 performance analyzer/profiler

I just wonder if anyone could suggest some good references for some hands on tutorials for NSight 3.0 on VS2010 (well, or some closest version)

The user doc is informative indeed but something more instructional will be good.

The GTC2013 Talk S3382 - Profiling and Optimizing CUDA Kernel Code with NVIDIA Nsight
Visual Studio Edition
describes how to use and interpret the Nsight VSE CUDA Profiler reports.

The talk focuses on CUDA Kernel Profiling. There is a small section at the beginning that covers CUDA trace. The PDF provides very little information as most of the session was a live demo so I recommend you watch the recording.