We investigate the Dio ́si-Penrose (DP) proposal for connecting the collapse of the wave function to gravity. The DP model needs a free parameter, acting as a cutoff to regularize the dynamics, and the predictions of the model highly depend on the value of this cutoff. The Compton wavelength of a nucleon seems to be the most reasonable cutoff value since it justifies the nonrelativistic approach. However, with this value, the DP model predicts an unrealistically high rate of energy increase. Thus, either one is forced to choose a much larger cutoff, which is not physically justified and totally arbitrary, or one needs to include dissipative effects in order to tame the energy increase. Taking the analogy with dissipative collisional decoherence seriously, we develop a dissipative generalization of the DP model. We show that even with dissipative effects, the DP model contradicts known physical facts, unless either the cutoff is kept artificially large or one limits the applicability of the model to massive systems. We also provide an estimation for the mass range of this applicability.
