Conjugate gradient heatbath for ill-conditioned actions

Abstract

We present a method for performing sampling from a Boltzmann distribution of an ill-conditioned quadratic action. This method is based on heatbath thermalization along a set of conjugate directions, generated via a conjugate-gradient procedure. The resulting scheme outperforms local updates for matrices with very high condition number, since it avoids the slowing down of modes with lower eigenvalue, and has some advantages over the global heatbath approach, compared to which it is more stable and allows for more freedom in devising case-specific optimizations.

References in corpus (6)

• Strict Detailed Balance is Unnecessary in Monte Carlo Simulation
• A New Approach to the Problem of Dynamical Quarks in Numerical Simulations of Lattice QCD
• Linear scaling electronic structure calculations and accurate sampling with noisy forces
• Stochastic linear scaling for metals and non metals
• Monte Carlo Quasi-Heatbath by approximate inversion
• The bermions: an approach to lattice QCD dynamical fermions from negative flavour numbers