Compile and install numpy and scipy linked with OpenBLAS. This makes them fast.
Currently supports:
- Ubuntu
Execute the following script
$ setup
and the fast numpy stack linked with OpenBLAS will be compiled and installed.
During the process, sudo is requested only for the installation of build
dependencies from the official repository. OpenBLAS, numpy and scipy will be
compiled from source and installed locally for the user. The OpenBLAS library
will be installed to $HOME/install
. The numpy and scipy libraries will be
installed with the --user
flag, which means that the packages will reside
in $HOME/.local/
.
Or manually, run the following in sequence (which is what the above setup
script does):
$ get_sources
$ make_openblas
$ make_numpy
$ make_scipy