Performance is crucial in scientific computing. There is a big difference if your experiments run one minute or one hour. We have already developed quite a bit of code, both in and outside packages, on which we are going to present some of the tooling that Julia provides for finding performance bottlenecks. Performance of your code or more precisely the speed of execution is of course relative (preference, expectation, existing code) and it's hard to find the exact threshold when we should start to care about it. When starting out with Julia, we recommend not to get bogged down by the performance side of things straightaway, but just design the code in the way that feels natural to you. As opposed to other languages Julia offers you to write the things "like you are used to" (depending on your background), e.g. for cycles are as fast as in C; vectorization of mathematical operators works the same or even better than in MATLAB, NumPy.
0 commit comments