3 ms·
True, but if performance is your key concern, you shouldn't use either of them. Octave and MATLAB are great for prototyping matrix based numerical algorithms. A
by aurelian15 10y ago
True, but if performance is your key concern, you shouldn't use either of them. Octave and MATLAB are great for prototyping matrix based numerical algorithms. And for this purpose I find Octave much more pleasant, since it is less nitpicky regarding its syntax. And once I have something that works and want to apply it to huge datasets, I usually rewrite my code in C++.
- IshKebab 10y agoThat's exactly why Julia was created - it's supposed to be like Matlab, but fast.