This approach would allow your profiler to provide the most information to the user, including capturing timings of all managed methods (regardless of the governing CLR), present interleaved call stacks including code from all runtimes, analyze all managed heaps from all runtimes, monitor the behavior of all managed code in the process, etc.