You need to install the following libraries: Eigen3, FFTW3, MathGL. For Ubuntu Linux just execute:

sudo apt install libeigen3-dev
sudo apt install libfftw3-dev
sudo apt install mathgl

More information about Eigen3 can be found here. More information about FFTW3 can be found here. More information about mathgl can be found here.

Also, it is recommended to install Intel TBB library for better performance. For Ubuntu Linux just execute:

sudo apt install libtbb-dev

More information about Intel TBB can be found here.

The easiest way to compile The GrPPI code is using an out-of-source directory under the GrPPI root directory:

mkdir grppi_workflows/build
cd grppi_workflows/build

Then, you may generate the scripts by just doing:

cmake ..

Afterwards, you launch the build process with:

make

Finally you copy the executable to the dispel4py tc_cross_correlation directory

cd ../..
cp grppi_workflows/build/use_cases/tc_cross_correlation/tc_cross_correlation dispel4py_workflows/tc_cross_correlation/

• The first workflow (realtime_prep.py) stores the results in a directory called OUTPUT/DATA.

• The second one (tc_cross_correlation) stores the results in OUTPUT/XCORR directory.

• This is our script for executing both workflows with multi mapping (dispel4py) and TBB mapping (GrPPI):

    export DISPEL4PY_XCORR_STARTTIME=2015-04-06T06:00:00.000
    export DISPEL4PY_XCORR_ENDTIME=2015-04-06T07:00:00.000
    rm -rf ./tc_cross_correlation/OUTPUT/DATA
    rm -rf ./tc_cross_correlation/OUTPUT/XCORR
    mkdir  ./tc_cross_correlation/OUTPUT/DATA
    mkdir ./tc_cross_correlation/OUTPUT/XCORR
  
    dispel4py multi tc_cross_correlation/realtime_prep.py -f tc_cross_correlation/realtime_xcorr_input.jsn -n 4
    tc_cross_correlation/tc_cross_correlation ./tc_cross_correlation/OUTPUT/ $DISPEL4PY_XCORR_STARTTIME tbb
  

• The last step is to open the file " tc_cross_correlation/realtime_xcorr_input.jsn” and change the path of the file Copy-Uniq-OpStationList-NetworkStation.txt

    xxx/tc_cross_correlation/Copy-Uniq-OpStationList-NetworkStation.txt  
  

• You could change the realtime_xcorr_input.jsn for using Uniq-OpStationList-NetworkStation.txt (it contains all the stations) instead of Copy-Uniq-OpStationList-NetworkStation.txt (it contains only a few of stations for testing the workflow). This data has been obtained from the IRIS website.

    {
    	"streamProducer" : [ { "input" : “/xxxxxxxxx/tc_cross_correlation/Uniq-OpStationList-NetworkStation.txt" } ]
    }
  

• It is important that you delete DATA and XCORR directories every time before starting to run your workflows.