Application: Gyro
Platform: Teragrid
Tool: HPC Toolkit
Usage: The application's batch job script needs to be modified by including the full path of "hpcrun" command before the application executable when giving these as parameters to mpirun.Eg.
Original application: mpirun -machinefile $PBS_NODEFILE -np 16 $GYRO_DIR/src/BigScience
To use HPC Toolkit: mpirun -machinefile $PBS_NODEFILE -np 16 /home/ac/sarats/tools/HPCToolkitRoot-1.1.0/HPCToolkit-ia64-Linux/bin/hpcrun $GYRO_DIR/src/BigScience
Each process emits it's own binary .0 file. Before you use hpcviewer to view the profiles, you need to run "hpcquick -I source_dir -P profile_files" to generate hpcquick database. Then after that is successfully completed, you can use hpcviewer to view the output.
Alternately, you can use hpcprof to create a XML file from the .0 files. Eg. hpcprof -p -e executable *.0 > hpcprof.out and then use hpcview.
Output( XML from hpcprof): B1-std | B3-gtc
Graphial output: Graphical output can be obtained using hpcview. I encountered some problems while parsing the configuration file before using hpcviewer. The issue has been traced down to the xerces xml parser with help from the HPC Toolkit team.(Thanks to Nathan Tallent and John Mellor-Crummey)