Senior R&D Manager (Data Science) at Synopsys Software Integrity Group and Treasurer at Farset Labs
My work has be flittering between Python and Matlab recently, and lets say I’m not a massive fan of Matlab at the best of time, and VM matlab isn’t the most performant thing in the world.
So I was happy to hear that
octave, an open source, Matlab compatible analysis framework have started testing their GUI.
sudo aptitude -y build-deps octave sudo aptitude -y install gfortran libgfortran3-dbg-arm64-cross liblapack-dev libblas-dev libarpack2-dev llvm-dev libfltk1.3-dev libglu1-mesa-dev libcurl4-gnutls-dev libfreetype6-dev libqt4-dev libfontconfig1-dev libfftw3-dev libqrupdate-dev libqscintilla2-designer texlive
/dev/shm as the build directory as it’s effectivly a RAM disk, but YMMV.
curl ftp://ftp.gnu.org/gnu/octave/octave-3.8.1.tar.gz | tar -xvzf - -C /dev/shm/; cd /dev/shm
I’m on a 64 bit machine with 4 real-cores, so I wanted openmp but do not enable 64bit addressing, it doesn’t work at the minute. Both of these features are experimental and YMMV. If someone comes up with a fix for 64bit, let me know.
Due to the way that
configure looks for the blas and lapack libraries, you have to tell it where to go.
./configure --enable-openmp --with-blas=/usr/lib/libblas/ --with-lapack=/usr/lib/liblapack/
May as well use all the cores for compilation
make -j4; make check
This is the only operation that needs
sudo make install
You’ll need to add
/usr/local/bin/ to your path in whichever fashion you prefer.
Enjoy some “tasty” Matlab style editing. If you break it it’s not my fault!