Optimizing off-lattice Diffusion-Limited Aggregation

We present a technique to improve the time scaling of Diffusion-Limited Aggregation simulations. The proposed method reduces the number of calculations by making an extensive use of the RAM memory to store information about the particles' positions and distances. We have simulated clusters up to 5⋅106 particles in 2D and up to 1⋅106 particles in 3D and compared the calculation times with previous algorithms proposed in the literature. Our method scales t∝Np1.08, outperforming the current optimization techniques.