I have two large 1-D arrays, T_A and T_B, that contain np.float64 and a rolling window of length m. Essentially, I am rol
T_A
T_B
np.float64
m