+ +
+

Installing HPAT

+

These commands install HPAT and its dependencies such as Numba, LLVM and HDF5 +on Ubuntu Linux:

+
$ sudo apt install llvm-4.0 make libc6-dev gcc-4.8
+$ # download and install Anaconda python distribution
+$ conda create -n HPAT
+$ source activate HPAT
+$ conda install numpy scipy pandas gcc mpich2 llvmlite
+$ git clone https://github.com/IntelLabs/numba.git
+$ cd numba
+$ git checkout hpat_req
+$ python setup.py develop
+$ cd ..
+$ # download hdf5 and cd inside
+$ CC=mpicc CXX=mpicxx ./configure --enable-parallel
+$ make; make install
+$ cd ..
+$ export HDF5_DIR=/home/user/hdf5-1.10.1/hdf5/
+$ export C_INCLUDE_PATH=$C_INCLUDE_PATH:$HDF5_DIR/include
+$ export CPLUS_INCLUDE_PATH=$CPLUS_INCLUDE_PATH:$HDF5_DIR/include
+$ export LD_LIBRARY_PATH=$LD_LIBRARY_PATH:$HDF5_DIR/lib
+$ export LIBRARY_PATH=$LIBRARY_PATH:$HDF5_DIR/lib
+$ git clone https://github.com/h5py/h5py.git
+$ cd h5py
+$ python setup.py configure --hdf5=$HDF5_DIR
+$ LDSHARED="mpicc -shared" CXX=mpicxx LD=mpicc CC="mpicc" python setup.py install
+$ cd ..
+$ git clone https://github.com/IntelLabs/hpat.git
+$ cd hpat
+$ LDSHARED="mpicxx -shared" CXX=mpicxx LD=mpicxx CC="mpicxx -std=c++11" python setup.py develop
+
+
+

Commands for running the logistic regression example:

+
$ python generate_data/gen_logistic_regression.py
+$ mpirun -n 2 python examples/logistic_regression.py
+
+
+
+ + +