Differences between regular and random order of updates in damage spreading simulations

TL;DR

This study examines how the order of site updates affects damage spreading in the 3D Ising model, revealing that random updates lead to larger damage and lower spreading temperature, indicating non-universality.

Contribution

It demonstrates that the update order significantly influences damage spreading behavior, a factor previously overlooked in Ising model studies.

Findings

Random update order results in larger stationary damage.

Spreading temperature is lower with random updates.

Damage spreading is non-universal across different update schemes.

Abstract

We investigate the spreading of damage in the three-dimensional Ising model by means of large-scale Monte-Carlo simulations. Within the Glauber dynamics we use different rules for the order in which the sites are updated. We find that the stationary damage values and the spreading temperature are different for different update order. In particular, random update order leads to larger damage and a lower spreading temperature than regular order. Consequently, damage spreading in the Ising model is non-universal not only with respect to different update algorithms (e.g. Glauber vs. heat-bath dynamics) as already known, but even with respect to the order of sites.