> mind bogglingly slow Python is outside of its optimised numerical science ecosystem Granted, but inside its optimised numerical science ecosystem, Python is

> mind bogglingly slow Python is outside of its optimised numerical science ecosystem Granted, but inside its optimised numerical science ecosystem, Python is, in fact, fast enough. If most of your program is calls into numpy, Python will get you where you need to go. In my experience, one scalar Python math operation takes about the same amount of time as the equivalent numpy operation on a million-element array. Linked against a recent libblas, numpy will even distribute work across multiple cores. So much for the GIL.