loopless: npix = 256, nbeams = 50, time = 2208.879 seconds I guess
loop-ful: npix = 256, nbeams = 50, time = 9086.7630867s
Efficient computing would actually use the CPU no? It doesn't appear to be very CPU intensive, so we should investigate why numpy isn't using all the cores.
Can probably be improved by clearing the unused arrays from memory. It's actually quite slow as-is
Gaussian Beamlets
Case: