The installation of SciPy from source would be straightforward if it weren’t for the additional libraries LAPACK and BLAS that need to be installed as well. While I’m not new to compiling packages from source and resolving dependencies, this one stumped me – I ended up with this error:

/usr/local/lib/libflapack.a(slaruv.o): relocation R_X86_64_32S against `a local symbol’ can not be used when making a shared object; recompile with -fPIC

That’s telling me that I need to recompile using the “-fPIC” option, which I had already done. Strange. Anyway, if you want to get this installed the easy way, stop trying to compile this from source and do this instead:

Open up Yast, and add this repository to your sources:

http://repos.opensuse.org/science/SUSE_Linux_10.1/

If you are using Suse 9.3, 10.0, or 10.2, simply change the “SUSE_Linux_10.1″ to “openSUSE_10.2″, “SUSE_Linux_10.0″, or “SUSE_Linux_9.3″ in the provided link. After you’ve added this repository, open “Software Management” and search for ‘SciPy’. You should find it. You can additionally search for ‘NumPy’, ‘Lapack’, and ‘blas’ (although the dependencies should be sorted out automatically).

I would recommend additionally installing ‘matplotlib’ for plotting. If you do that, you’ll need to grab the matplotlibrc file and stick it in your .matplotlib directory. The only option I had to change to get plotting to work was the backend – I chose “QtAgg”.

Thats it! When trying to set this up, I found lots of Suse users with installation/compilation issues – but nobody seemed to know this simple method of installation.