How to make ATLAS to be Avaliable for Octave?

Question

I have been using octave for some Machine Learning algorithms, but when using Neural Networks, it uses the 100% of 1 core, I would like it to use all my 4 cores, I read this could be done with atlas, but when I tryed to install it on ubuntu (13), I noticed it was already installed; Nevertheless, it looks that octave can't see it

Best Regards!

Answer 1

One possible way of increasing the usage of CPU cores is vectorizing your Neural Network implementation.

More information about vectorization can be found in the following tutorial.

http://ufldl.stanford.edu/wiki/index.php/Vectorization