Monte Carlo Profiling (Windows Embedded CE 6.0)
In Monte Carlo profiling, the kernel profiler interrupts the OS periodically and samples the program counter for the CPU. The kernel profiler then correlates each sample with the function that the samples is inside, and totals up the number of samples per function. The kernel profiler then generates a report of the modules and functions with the most profiler hits.
By interrupting the OS periodically, you can learn which functions run most of the time. You can monitor OS processes, such as Nk.exe, and user processes.
Monte Carlo profiling requires that the OAL implement an interrupt for the kernel profiler. For more information, see Setting Up Monte Carlo Profiling.
You can perform Monte Carlo profiling with or without a buffer that holds samples collected by the kernel profiler. For more information about the benefits and disadvantages of buffered and unbuffered profiling, see Buffered and Unbuffered Profiling.
Monte Carlo profiling does not measure or record execution time.