The quality of sound-field reproduction is strongly dependent on the placement of the secondary sources used to generate the desired sound-field. In this communication, we focus on the case of control by matching the generated sound-field with the desired values at particular points in the reproduction zone. Several loudspeaker and control point placements are compared, such as regular placements, FrameSense, Empirical Interpolation Method, etc. Numerical results show that regular placements are (relatively) inefficient, control points located inside the reproduction zone are necessary, and that best results are obtained by jointly optimizing the positions of the sources and control points with the Empirical Interpolation Method, which also minimizes the required output power of the loudspeakers. This method is also shown to be robust with respect to moderate position errors of the sources and control points.