Profiling CESM with TAU

[nitkbhat@gmail_com]

I recently used TAU to profile the Community Earth System Model (CESM). I am running this on 128 cores spread across 8 nodes. 
On completion of the execution I get 128 profiles (each for the 128 processes). When I load these profiles using the paraprof, I get the graph suggesting metrics.  However, I am only able to get the data for MPI based functions ( like MPI gather, MPI scatter, broadcast) and not for the user defined routines and calls. I do not want the profiling to be carried out for these MPI functions. I want it to be done for the user defined fortran subroutines existing in the code. So that, I can identify the user defined functions which take the most amount of time.  I have used other tools like (hpctoolkit, ITAC, Vtune, Allinea Map etc). But, I am keen on getting the time for each function call. for that, I want to get the number of function calls made for each function. TAU (using instrumentation) provides these metrics to me. But, I am not able to profile the function times and number of calls for the user defined F90 functions. 
Is there any help I can get regarding profiling CESM with TAU? 

Thanks,Nitin K Bhat,Supercomputer Education and Reasearch Center,Indian Institute of Science