Low-cost bitwise 2D ising model representation using Monte Carlo and metropolis algorithm

摘要

The 2D Ising model obviously deals with mass data to study the behavior of magnetization and energy related with temperature. This simulation should be worked out efficiently in terms of cost, time and memory consumption. In this paper, we introduce an innovative technique that executes bits instead of integers in order to reduce memory usage (1/32) and turnaround time (0.53). This approach has been implemented in two ways using C++: 1) Boolean 1D-array lattice representation, 2) Individual bits of unsigned integers. Improvement on the spatiotemporal performance of this approach over the method of traditional integer representation is proved both analytically and experimentally.