Tutorial

pprof

pprof is a parallel profile data display tool. It can be used to display the function and aggregate collection data for pC++, HPC++ and POOMA applications. It supports static and dynamic profiling displays. Pooma applications use dynamic profiling.

Using pprof 


usage: pprof [-c|-b|-m|-t|-e|-i] [-r] [-s] [-n num] [-f filename] [-l] [node numbers]
 -c : Sort according to number of Calls
 -b : Sort according to number of suBroutines called by a function
 -m : Sort according to Milliseconds (exclusive time total)
 -t : Sort according to Total milliseconds (inclusive time total) (DEFAULT)
 -e : Sort according to Exclusive time per call (msec/call)
 -i : Sort according to Inclusive time per call (total msec/call)
 -v : Sort according to standard deViation (excl usec)
 -r : Reverse sorting order
 -s : print only Summary profile information
 -n num : print only first  number of functions
 -f filename : specify full path and Filename without node ids 
 -l : List all functions and exit 
[node numbers] : prints only info about all contexts/threads of given node numbers

for e.g.,
To display the profile data sorted by exclusive time spent per call and the data is in the current working directory, use 
% pprof -e
To just display the names of all functions in profile files in directory ./data use 
% pprof -l -f ./data/profile
[pprof

The above figure shows pprof being used with a POOMA application highlighting PETE functions.