On Tue, Feb 27, 2018 at 9:02 AM, Seb <splu...@gmail.com> wrote: > That's right. I just tried this manipulation by replacing the last > block of code in my example, from the line above `for` loop with: > > ---<--------------------cut here---------------start------------------->--- > # Alternative using `np.matmul` > uvw_alt = uvw.reshape((uvw.shape[0], 1, uvw.shape[1])) > bmats = np.asarray(map(randint_mat, maxint)) > uvw_rots_alt = np.matmul(uvw_alt, bmats).squeeze() > ---<--------------------cut here---------------end--------------------->--- > > Interestingly, the time savings from IPython are not spectacular:
The second line is still calling the Python function randint_mat 1000 times. That should be possible to do without the Python function using broadcasting, if https://github.com/numpy/numpy/issues/6745 were fixed. Otherwise, I don't know of a way to vectorize that part efficiently. -- https://mail.python.org/mailman/listinfo/python-list
