arXiv subjects

Stephan Durr

Publications and source records attributed to Stephan Durr.

At least 37 records · Page 2Linked to original sources

arXiv2018

Optimization of the Brillouin operator on the KNL architecture

Experiences with optimizing the matrix-times-vector application of the Brillouin operator on the Intel KNL processor are reported. Without adjustments to the memory layout, performance figures of 360 Gflop/s in single and 270 Gflop/s in double precision are observed. This is with N_c=3 colors, N_v=12 right-hand-sides, N_{thr}=256 threads, on lattices of size 32^3*64, using exclusively OMP pragmas. Interestingly, the same routine performs quite well on Intel Core i7 architectures, too. Some observations on the much harder Wilson fermion matrix-times-vector optimization problem are added.

hep-lat↗

arXiv2018

User's guide to Monte Carlo methods for evaluating path integrals

We give an introduction to the calculation of path integrals on a lattice, with the quantum harmonic oscillator as an example. In addition to providing an explicit computational setup and corresponding pseudocode, we pay particular attention to the existence of autocorrelations and the calculation of reliable errors. The over-relaxation technique is presented as a way to counter strong autocorrelations. The simulation methods can be extended to compute observables for path integrals in other settings.

physics.comp-ph↗

arXiv2017

On the suitability of the Brillouin action as a kernel to the overlap procedure

We investigate the Brillouin action in terms of its suitability as a kernel to the overlap procedure, with a view on both heavy and light quark physics. We use the diagonal elements of the Kenney-Laub family of iterations for the sparse matrix sign function, since they grow monotonically and facilitate cascaded preconditioning strategies with different rational approximations to the sign function. We find that the overlap action with the Brillouin kernel is significantly better localized than the version with the Wilson kernel.

hep-lat↗

arXiv2016

First numerical experiences with overlap fermions based on the Brillouin kernel

Numerical experiences are reported with overlap fermions which employ the Brillouin action as a kernel. After discussing the dispersion relations of both the kernel and the resulting chiral action, some of the physics features are addressed on quenched backgrounds. We find that the overlap with Brillouin kernel is much better localized than the overlap with Wilson kernel. Also a preliminary account is given of the cost of the formulation, in terms of CPU time and memory.

hep-lat↗

arXiv2016

Leptonic decay-constant ratio $f_K/f_π$ from clover-improved $N_f=2+1$ QCD

The leptonic decay-constant ratio $f_K/f_π$ is calculated from lattice-QCD simulations using $N_f=2+1$ dynamical fermion flavors in the clover-improved formulation and 2-HEX smearing. The simulations were performed for a range of mass-degenerate light quarks including the physical point and at various lattice couplings and volumes, allowing to quantify all relevant sources of systematic uncertainties for our final number of the decay-constant ratio. Utilizing input from chiral perturbation theory, we also quote the charged decay-constant ratio $f_{K^\pm}/f_{π^\pm}$. With further input from super-allowed nuclear $β$-decays, eventually we obtain an estimate for the CKM-matrix element $V_{\rm us}$.

hep-lat↗

arXiv2015

QCD thermodynamics with continuum extrapolated Wilson fermions II

We continue our investigation of 2+1 flavor QCD thermodynamics using dynamical Wilson fermions in the fixed scale approach. Two additional pion masses, approximately 440 MeV and 285 MeV, are added to our previous work at 545 MeV. The simulations were performed at 3 or 4 lattice spacings at each pion mass. The renormalized chiral condensate, strange quark number susceptibility and Polyakov loop is obtained as a function of the temperature and we observe a decrease in the light chiral pseudo-critical temperature as the pion mass is lowered while the pseudo-critical temperature associated with the strange quark number susceptibility or the Polyakov loop is only mildly sensitive to the pion mass. These findings are in agreement with previous continuum results obtained in the staggered formulation.

hep-lat↗

arXiv2015

$D_s \rightarrow η, η'$ semileptonic decay form factors with disconnected quark loop contributions

We calculate for the first time the form factors of the semi-leptonic decays of the $D_s$ meson to $η$ and $η^\prime$ using lattice techniques. As a by-product of the calculation we obtain the masses and leading distribution amplitudes of the $η$ and $η^\prime$ mesons. We use $N_f=2+1$ non-perturbatively improved clover fermions on configurations with a lattice spacing $a\sim 0.075$ fm. We are able to obtain clear signals for relevant matrix elements, using several noise reduction techniques, both for the connected and disconnected contributions. This includes a new method for reducing the variance of pseudoscalar disconnected two-point functions. At zero momentum transfer, we obtain for the scalar form factors, $|f_0^{D_s\to η}|=0.564(11)$ and $|f_0^{D_s\to η'}|=0.437(18)$ at $M_π\approx 470\, {\rm MeV}$, as well as $|f_0^{D_s\to η}|=0.542(13)$ and $|f_0^{D_s\to η'}|=0.404(25)$ at $M_π\approx 370\, {\rm MeV}$, where the errors are statistical only.

hep-lat↗

arXiv2014

Validity of ChPT -- is M_π=135 MeV small enough ?

I discuss the practical convergence of the SU(2) ChPT series in the meson sector, based on 2+1 flavor lattice data by the Wuppertal-Budapest and Budapest-Marseille-Wuppertal collaborations. These studies employ staggered and clover-improved Wilson fermions, respectively. In both cases large box volumes and several lattice spacings are used, and the pion masses reach down to the physical mass point. We conclude that LO and NLO low-energy constants can be determined with controlled systematics, if there is sufficient data between the physical mass point and about 350 MeV pion mass. Exploratory LO+NLO+NNLO fits with a wider range reveal some distress of the chiral series near M_π~ 400 MeV and suggest a complete breakdown beyond M_π~ 500 MeV.

hep-lat↗

arXiv2014

Charmonium spectral functions from 2+1 flavour lattice QCD

Finite temperature charmonium spectral functions in the pseudoscalar and vector channels are studied in lattice QCD with 2+1 flavours of dynamical Wilson quarks, on fine isotropic lattices (with a lattice spacing of 0.057 fm), with a non-physical pion mass of $m_π \approx$ 545 MeV. The highest temperature studied is approximately $1.4 T_c$. Up to this temperature no significant variation of the spectral function is seen in the pseudoscalar channel. The vector channel shows some temperature dependence, which seems to be consistent with a temperature dependent low frequency peak related to heavy quark transport, plus a temperature independent term at ω>0. These results are in accord with previous calculations using the quenched approximation.

hep-lat↗

arXiv2013

Lattice QCD at the physical point meets SU(2) chiral perturbation theory

We perform a detailed, fully-correlated study of the chiral behavior of the pion mass and decay constant, based on 2+1 flavor lattice QCD simulations. These calculations are implemented using tree-level, O(a)-improved Wilson fermions, at four values of the lattice spacing down to 0.054 fm and all the way down to below the physical value of the pion mass. They allow a sharp comparison with the predictions of SU(2) chiral perturbation theory (χPT) and a determination of some of its low energy constants. In particular, we systematically explore the range of applicability of NLO SU(2) χPT in two different expansions: the first in quark mass (x-expansion), and the second in pion mass (ξ-expansion). We find that these expansions begin showing signs of failure around M_π=300 MeV for the typical percent-level precision of our N_f=2+1 lattice results. We further determine the LO low energy constants (LECs), F=88.0 \pm 1.3\pm 0.3 and B^\msbar(2 GeV)=2.58 \pm 0.07 \pm 0.02 GeV, and the related quark condensate, Σ^\msbar(2 GeV)=(271\pm 4\pm 1 MeV)^3, as well as the NLO ones, l_3=2.5 \pm 0.5 \pm 0.4 and l_4=3.8 \pm 0.4 \pm 0.2, with fully controlled uncertainties. We also explore the NNLO expansions and the values of NNLO LECs. In addition, we show that the lattice results favor the presence of chiral logarithms. We further demonstrate how the absence of lattice results with pion masses below 200 MeV can lead to misleading results and conclusions. Our calculations allow a fully controlled, ab initio determination of the pion decay constant with a total 1% error, which is in excellent agreement with experiment.

hep-lat↗

arXiv2013

SU(2) chiral perturbation theory low-energy constants from 2+1 flavor staggered lattice simulations

We extract the next-to-leading-order low-energy constants \bar\ell_3 and \bar\ell_4 of SU(2) chiral perturbation theory, based on precise lattice data for the pion mass and decay constant on ensembles generated by the Wuppertal-Budapest Collaboration for QCD thermodynamics. These ensembles feature 2+1 flavors of two-fold stout-smeared dynamical staggered fermions combined with Symanzik glue, with pion masses varying from 135 to 435 MeV, lattice scales between 0.7 and 2.0 GeV, while m_s is kept fixed at its physical value. Moderate taste splittings and the scale being set through the pion decay constant allow us to restrict ourselves to the taste pseudoscalar state and to use formulas from continuum chiral perturbation theory. Finally, by dropping the data points near 135 MeV from the fits, we can explore the range of pion masses that is needed in SU(2) chiral perturbation theory to reliably extrapolate to the physical point.

hep-lat↗

arXiv2013

Taste-split staggered actions: eigenvalues, chiralities and Symanzik improvement

The eigenvalue spectra of staggered fermions with an Adams and/or Hoelbling mass term are studied. The chiralities of the eigenmodes reflect whether the chirality linked to the unflavored approximate (γ_5 \times 1) or the flavored exact (γ_5 \times ξ_5) staggered symmetry is considered, and which one of the RR, LR, RL, LL eigenmode definitions is used. In either case a sensitivity to the topological charge of the gauge background is found. We discuss how to remove the leading cut-off effects of these actions by means of a properly tuned improvement term and/or the overlap procedure. The combination of Symanzik improvement and link smearing radically improves the properties of the physical branch.

hep-lat↗

arXiv2013

Convergence issues in ChPT: a lattice perspective

This review addresses the practical convergence of the ChPT series in the p-regime. In the SU(2) framework there is a number of new results, and improved estimates of \bar\ell_3 and \bar\ell_4 are available. In the SU(3) framework few new lattice computations have appeared and the improvement in the precision of the low-energy constants L_i is comparatively slow. I sketch some of the convergence issues genuine to extensions of ChPT which include additional sources of chiral symmetry breaking (finite lattice spacing) and/or violations of unitarity (different sea and valence quark masses). Finally, it is pointed out that the quark mass ratios m_u/m_d, m_s/m_d happen to be such that no reordering of the chiral series is needed to accommodate the experimental pion and kaon masses.

hep-lat↗

arXiv2013

Meson and Baryon dispersion relations with Brillouin fermions

We study the dispersion relations of mesons and baryons built from Brillouin quarks on one N_f=2 gauge ensemble provided by QCDSF. For quark masses up to the physical strange quark mass, there is hardly any improvement over the Wilson discretization, if either action is link-smeared and tree-level clover improved. For quark masses in the range of the physical charm quark mass, the Brillouin action still shows a perfect relativistic behavior, while the Wilson action induces severe cut-off effects. As an application we determine the masses of the Ω_c^0, Ω_{cc}^+ and Ω_{ccc}^{++} baryons on that ensemble.

hep-lat↗

arXiv2013

Determination of SU(2) ChPT LECs from 2+1 flavor staggered lattice simulations

By fitting pion masses and decay constants from 2+1 flavor staggered lattice simulations to the predictions of NLO and NNLO SU(2) chiral perturbation theory we determine the low-energy constants l_3 and l_4. The lattice ensembles were generated by the Wuppertal-Budapest collaboration and cover pion masses in the range of 135 to 435 MeV and lattice scales between 0.7 and 2.0 GeV. By choosing a suitable scaling trajectory, we were able to demonstrate that precise and stable results for the LECs can be obtained from continuum ChPT to NLO. The pion masses available in this work also allow us to study the applicability of using ChPT to extrapolate from higher masses to the physical pion mass.

hep-lat↗

arXiv2013

Recent Progress in Lattice QCD

Recent progress in Lattice QCD is highlighted. After a brief introduction to the methodology of lattice computations the presentation focuses on three main topics: Hadron Spectroscopy, Hadron Structure and Lattice Flavor Physics. In each case a summary of recent computations of selected quantities is provided.

hep-lat↗

arXiv2012

Physics of eta-prime with rooted staggered quarks

The quark-mass dependence of the eta in the Schwinger model, which -- like the eta-prime in QCD -- becomes massive through the axial anomaly, is studied on the lattice with N_f=0,1,2. Staggered quarks are used, with a rooted determinant for N_f=1. In the chiral limit the Schwinger mass is reproduced, which suggests that the anomaly is being treated correctly.

hep-lat↗

arXiv2012

Brillouin improvement for Wilson fermions

We present a parameter-free Wilson-type lattice Dirac operator with an 81-point stencil for the covariant derivative and the Laplacian which attempts to minimize the breaking of rotational symmetry near the boundary of the Brillouin zone. The usefulness of this "Brillouin operator" in practical applications is explored by studying the scaling of pseudoscalar decay constants in quenched QCD, with rather good results in the physical charm region. We also investigate the suitability of this operator as a kernel to the overlap procedure. Here, the resulting overlap operator is found to be cheaper to construct and significantly better localized than the variety with the standard Wilson kernel.

hep-lat↗