Simulation of magnetic resonance static powder spectra is performed by a (possibly weighted) summation of single-crystal spectra computed for different orientations of the external field with respect to the principal axes of the magnetic interactions. The many available methods differ in the choice of the integration points (i.e., orientations) and weights, the set of which is called spherical code. There is continuing interest in minimizing the number of integration points necessary to a good simulation. Neglecting the possible interpolation of transition frequencies and intensities between integration points, we turn our attention to the efficiency of spherical codes themselves. To this end, an unbiased quantitative procedure to assess their efficiency in simulating magnetic resonance static powder spectra is proposed. To achieve an impartial judgement, the procedure has been designed by carefully taking into consideration the following points: choice of exact reference spectra; accurate definition of the merit figures; extended range of number of integration points; orientation dependence of the efficiency. The proposed procedure has been applied to an inclusive set of 23 spherical codes. It was found that most codes perform rather similarly. SPIRAL is the most efficient code, whereas Monte Carlo and "repulsive" codes show the best rotational invariance of the simulated lineshape with respect to the orientation of the spherical code.