QCD with light Wilson quarks on fine lattices (II): DD-HMC simulations and data analysis

Abstract

In this second report on our recent numerical simulations of two-flavour QCD, we provide further technical details on the simulations and describe the methods we used to extract the meson masses and decay constants from the generated ensembles of gauge fields. Among the topics covered are the choice of the DD-HMC parameters, the issue of stability, autocorrelations and the statistical error analysis. Extensive data tables are included as well as a short discussion of the quark-mass dependence in partially quenched QCD, supplementing the physics analysis that was presented in the first paper in this series.

Spectral density of the Hermitean Wilson Dirac operator: a NLO computation in chiral perturbation theory

“…For N f = 2 unimproved Wilson fermions, empirical observations [6] indicate that the median of the spectral gap distribution is linearly proportional to the quark mass m, while the width is basically independent on m and scales like ∼ a/ √ V , where a is the lattice spacing. On the other hand, for the O(a)-improved theory, the situation is less clear [8] and those properties have not been confirmed. In [7] the mode number of the O(a)-improved Wilson Dirac operator is computed, finding a nearly linear behavior up to 100 MeV above the threshold.…”

“…As suggested in [34], we perform an analytic continuation for the ghost components 8) JHEP04 (2011)031 where nowφ 5,6 are real variables taking values along the real axis. The LO potential can be separated in quark and ghost components…”

Recent progress on chiral symmetry breaking in QCD

“…collaboration 5 [48,15]. The parameter ranges that they span are a = 0.048-0.075, m = 6-37 MeV, and Λ = 20-500 MeV.…”

Confining force and running coupling with twelve fundamental and two sextet fermions

“…From the Wilson loops, we extract the effective mass V (r,t) = − lnW (r,t + 1)/W (r,t), which at sufficiently large times can be fitted to a constant V (r), including the covariance matrix for the correlated data. We use the double (nested) jackknife method for the error analysis: for every outer jackknife sample of Wilson loops, we have an inner jackknife loop to determine the covariance matrix for the effective masses, which is included in the fit [10]. We bin the data set until we typically have a total of between 10 and 20 bins.…”