<scp>quantizer</scp>: a new collisionless <i>N</i>-body code optimized for small-scale devices

ABSTRACT We present quantizer, a new collisionless N-body code tailored for cosmological and galaxy-scale simulations. The code uses the particle-mesh method, combined with a novel adaptive mesh refinement (AMR) technique. The design of the AMR method of quantizer maximizes the efficient use of refinement zones by combining a cell-by-cell scheme with parallelization-friendly regular patches. We test the code using three astrophysical systems: a binary system of two particles without AMR to test orbital accuracy; an isolated elliptical galaxy modelled as a Plummer sphere; and a merger of two disc galaxies. Comparisons are made with the well-known particle-mesh AMR code ramses, and with our own direct particle-particle solver. The results show that the accuracy of quantizer is equal to or better than ramses, and the runtime is approximately 2.5 to 10 times faster than that code, depending on the problem.