This repository contain an implementation of the Iterative Proportional Fitting Procedure (IPFP) which can be run on large HPC cluster. This implementation is thought to speed-up the the IPFP step used by Albani et al., where they use this algorithm to forecast how many people will frequent a point of interest (eg. restaurants, shops, ...) in a given hour of the day.
- ...
needs testing
To compile from source execute:
cd ipfp-hpc-cluster
mkdir build
cd build
cmake -DMPI_CXX_COMPILER=/usr/local/bin/mpicxx -DMPI_C_COMPILER=/usr/local/bin/mpicc -DMPIEXEC_EXECUTABLE=/usr/local/bin/mpiexec.hydra ..
make -j $(nproc)