Simulations of gas-solid mixtures are used in many scientific and industrial applications. Two-Fluid Smoothed Particle Hydrodynamics (TFSPH) is an approach when gas and solids are simulated with different sets of particles interacting via drag force. Several methods are developed for computing drag force between gas and solid grains for TFSPH. Computationally challenging are simulations of gas-dust mixtures with intense intephase interaction, when velocity relaxation time tstop is much smaller than dynamical time of the problem. In explicit schemes the time step t must be less than tstop, that leads to high computational costs. Moreover, it is known that for stiff problems both grid-based and particle methods may require unaffordably detailed resolution to capture the asymptotical bahaiviour of the solution. To address this problem we developed fast and robust method for computing stiff and mild drag force in gas solid-mixtures based on the ideas of Particle-in-Cell approach. In the paper we compare the results of new and previously developed methods on test problems.