3 ms·
Julia interop with Python is insanely easy. using PyCall np = pyimport("numpy") res = np.fft.fft(rand(ComplexF64, 10)) You just called numpy fft from Julia.
by adamnemecek 6y ago
Julia interop with Python is insanely easy.
using PyCall
np = pyimport("numpy")
res = np.fft.fft(rand(ComplexF64, 10))
You just called numpy fft from Julia.
- kwertzzz 6y agoYes, and the interface is pretty fast and without any noticeable overhead: julia> data = rand(ComplexF64, 1024^2); # python fft from julia: julia> res = @btime np.fft.fft(data); 78.613 ms (39 allocations: 16.00 MiB) # python fft in ipython: In [11]: %timeit res = np.fft.fft(data) 89.3 ms ± 1.65 ms per loop (mean ± std. dev. of 7 runs, 10 loops each) As expected, julia has its own fft package (based on FFTW): julia> res = @btime fft(data); 61.540 ms (33 allocations: 16.00 MiB)