Speeding up the hybrid Monte Carlo algorithm for dynamical fermions

By M. Hasenbusch. 2001

303 citations recorded by the source corpus.

Abstract We propose a modification of the hybrid Monte Carlo algorithm that allows for a larger step-size of the integration scheme at constant acceptance rate. The key ingredient is that the pseudo-fermion action is split into two parts. We test our proposal at the example of the two-dimensional lattice Schwinger model with two degenerate flavours of Wilson-fermions.

Subfields: Nuclear and High Energy Physics

View Paper: https://www.semanticscholar.org/paper/7d0e20713e509c025fff1fcb1ee5ff1e71e96c1b

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Nuclear and High Energy Physics

Citations in the corpus

Light Hadron Masses from Lattice QCD (2012, background): “Splitting off the UV modes of the spectrum by mass preconditioning (Hasenbusch, 2001; Hasenbusch and Jansen, 2003) or via domain decomposition (Luscher, 2003, 2004, 2005) and integrating IR and UV part with different time steps leads to a substantial additional speedup.”
Nonperturbative QCD Simulations with 2+1 Flavors of Improved Staggered Quarks (2009, methodology): “We used four sets ofpseudofermion fields and corresponding rational functions (Hasenbusch, 2001; Hasenbusch and Jansen, 2003).”
The strange quark mass and Lambda parameter of two flavor QCD (2012, methodology): “and then apply two levels of mass preconditioning [21] to its Schur complement QS = 1−Q−1 ee QeoQ ooQoe , i.”
theta dependence of SU(N) gauge theories in the presence of a topological term (2007, methodology): “symmetry. Indeed, it is expected to substantially worsen the cost estimates of the dynamical fermion simulations for lattice QCD using the actual optimal hybrid Monte Carlo algorithms, see e.g. Refs. [220,297, 337,386,442]. Note that standard cost estimates of these algorithms, see e.g. Refs. [84,18”
Higher representations on the lattice: numerical simulations. SU(2) with adjoint fermions (2008, background): “Introduction Recent algorithmic progress in simulations of QCD with dynamical fermions have shown that current computing facilities can reach beyond the quenched approximation without compromising the robustness of the formulation [1, 2, 3, 4, 5, 6].”
Optimal tuning of the hybrid Monte Carlo algorithm (2010, unknown): “It has since become popular in a number of application areas, including statistical physics [1,17,18,23,43], computational chemistry [22,24,30,42,45], data assimilation [2], geophysics [30] and neural networks [32,46].”
Transport properties of the quark-gluon plasma (2011, methodology): “ral pseudofermions (instead of just one), where the long wavelength modes are separated from short wavelength modes. The most widely used realizations of this idea are the mass preconditioning (MPHMC [117]), the domain-decomposition (DDHMC [118]) and the rational HMC (RHMC [119]) algorithms. The cal”
Simulation of QCD with Nf = 2 + 1 flavors of non-perturbatively improved Wilson fermions (2014, methodology): “To this end we use Hasenbusch’s mass factorization [36] with a twisted mass [37] applied to the last term in eq.”