Since the generated jit code could have been generated long before the profiling starts,
but be used during profiling, it is probably important to ensure that the mechanism for
identifying the jitted code allows for persistent data. We choose files with names that
included owning PIDs to provide this information for multiple sources (multiple jvm's).
If you wish to provide for a mechanism that allows for reuse of generated code areas,
then time stamps are probably indicated. Time stamps for the generated code as well
as time stamps for anonymous ticks, this can be used to map the tick to a jitted method.
If you wish to further close address to name holes, and correlate the time stamps,
consider SMPs that allow for out of sync time stamps. Then one would also need
to know for which processor the time stamps are applicable .