4 ms·
Probably not faster in an absolute sense but things like loops in Julia can be properly optimized and will sometimes be more readable than structuring your prog
by jolux 5y ago
Probably not faster in an absolute sense but things like loops in Julia can be properly optimized and will sometimes be more readable than structuring your program entirely around NumPy constructs.
- longqzh 5y agoFor numpy, it's correct. But numba can optimize the loop. So optimized and readable loop is not an advantage of julia compare to numba.
- ChrisRackauckas 5y agoNot in real-world contexts. This is spelled out in Julia for Biologists (https://arxiv.org/abs/2109.09973 https://arxiv.org/abs/2109.09973) which does the operation counting to show why using Numba with SciPy is still an order of magnitude slower in scientific operations like solving differential equations compared to Julia. An order of magnitude on widely used scientific analyses is pretty significant!