This example can be used on one or more cores using MPI, pthread, OpenMP or
with a hybrid MPI+<thread> runtime. It features a selective instrumentation
file that uses outer-loop level instrumentation in two routines. 

The tau.conf configuration file contains the environment variable settings
necessary to configure TAU at runtime to dump the current timer stack state
and timer values for the timers on the state.

The signalling script is used to launch the program, get its process id, and
periodically signal the process to dump the callstack. The callstack values
are dumped to callstack.* files, indexed by process index. All threads within
the process, including GPU threads, are dumped to the same file.

