I'd like you all to be aware of an issue that can affect the quality or functionality of the Concurrency Visualizer. Our tool relies heavily on gathering timestamps in order to correlate events across cores and threads. When running on a virtualized processor, depending on many variables, the fidelity of this information can be affected when running on a Hyper-V system. The Concurrency Visualizer can even fail when it detects inconsistent timing information. We are working on addressing this in future versions, but for now, we recommend that you do your performance analysis running on native operating systems. Some purists in the performance engineering community might even claim that doing performance analysis on a hyper-V is fundamentally wrong, just like doing performance analysis with a lot of interference on a system from external applications or sources. I'm not such a purist on the former because VM environments will continue to increase in popularity and we need a good solution; however, for now you should be warned against it.
We apologize for any inconvenience that this may cause you. If you have any experiences that you'd like to share in this regard, please don't hesitate to do so here.