arXiv subjects
Publications and source records attributed to R. Frezzotti.
The phase structure of zero temperature twisted mass lattice QCD is investigated. We find strong metastabilities in the plaquette observable when the untwisted quark mass assumes positive or negative values. We provide interpretations of this phenomenon in terms of chiral symmetry breaking and the effective potential model of Sharpe and Singleton.
In this paper we discuss how the peculiar properties of twisted lattice QCD at maximal twist can be employed to set up a consistent computational scheme in which, despite the explicit breaking of chiral symmetry induced by the presence of the Wilson and mass terms in the action, it is possible to completely bypass the problem of wrong chirality and parity mixings in the computation of the CP-conserving matrix elements of the $ΔS=1,2$ effective weak Hamiltonian and at the same time have a positive determinant for non-degenerate quarks as well as full O($a$) improvement in on-shell quantities with no need of improving the lattice action and the operators.
The phase structure of zero temperature twisted mass lattice QCD is investigated. We find strong metastabilities in the plaquette observable when the untwisted quark mass sweeps across zero.
We show that it is possible to improve the chiral behaviour and the approach to the continuum limit of correlation functions in lattice QCD with Wilson fermions by taking arithmetic averages of correlators computed in theories regularized with Wilson terms of opposite sign. Improved hadronic masses and matrix elements can be obtained by similarly averaging the corresponding physical quantities separately computed within the two regularizations. To deal with the problems related to the spectrum of the Wilson--Dirac operator, which are particularly worrisome when Wilson and mass terms are such as to give contributions of opposite sign to the real part of the eigenvalues, we propose to use twisted-mass lattice QCD for the actual computation of the quantities taking part to the averages. The choice $\pm π/2$ for the twisting angle is particularly interesting, as O($a$) improved estimates of physical quantities can be obtained even without averaging data from lattice formulations with opposite Wilson terms. In all cases little or no extra computing power is necessary, compared to simulations with standard Wilson fermions or twisted-mass lattice QCD.
The maximally twisted lattice QCD action of an $SU_f(2)$ doublet of mass degenerate Wilson quarks gives rise to a real positive fermion determinant and it is invariant under the product of standard parity times the change of sign of the coefficient of the Wilson term. The existence of this spurionic symmetry implies that O($a$) improvement is either automatic or achieved through simple linear combinations of quantities taken with opposite external three-momenta. We show that in the case of maximal twist all these nice results can be extended to the more interesting case of a mass non-degenerate quark pair.
It is possible to improve the chiral behaviour and the approach to the continuum limit of correlation functions in lattice QCD with standard and twisted Wilson fermions by taking arithmetic averages of correlators computed in theories regularized with Wilson terms of opposite sign. To avoid the problem of the spurious zero-modes of the Wilson--Dirac operator, twisted-mass lattice QCD should be used for the actual computation of the correlators taking part in the averages. A particularly useful choice for the twisting angle is $\pmπ/2$ where many physical quantities (e.g. masses and zero-momentum matrix elements) are automatically improved with no need of averaging.
Lattice formulations of QCD with Wilson fermions and a chirally twisted quark mass matrix provide an attractive framework for non-perturbative numerical studies. Owing to reparameterization invariance, the limiting continuum theory is just QCD. No spurious quark zero modes, which are responsible for the problem with exceptional configurations, can occur at finite values of the quark mass. Moreover, the details of the lattice formulation can be adjusted so as to simplify the renormalization and the O($a$) improvement of several quantities of phenomenological relevance. The first exploratory studies in the quenched approximation yield very encouraging results.
We report on the latest results on the running coupling of two flavour QCD in the Schrödinger functional scheme. Results for the step scaling function are obtained from simulations on lattices $L/a=8$ and $L/a=16$ which confirm the first results from lattices $L/a=4,5,6$ presented one year ago by the ALPHA collaboration. We also discuss some algorithmic aspects, in particular concerning the occurrence of metastable states. A modified sampling, in order to estimate the proper weight of these states in the path integral, is proposed and tested.
We present a first numerical study of lattice QCD with O(a) improved Wilson quarks and a chirally twisted mass term. Renormalized correlation functions are derived from the Schroedinger functional and evaluated in an intermediate space-time volume of size 0.75^3 x 1.5 fm^4. In the quenched approximation precise results are then obtained with a moderate computational effort, allowing for a detailed study of the continuum approach. The latter is discussed in terms of observables which converge to meson masses and decay constants in the limit of large space-time volume. In the O(a) improved theory we find residual cutoff effects to be at the level of a few percent for lattice spacings of about 0.1 fm.
APENEXT is a new generation APE processor, optimized for LGT simulations. The project follows the basic ideas of previous APE machines and develops simple and cheap parallel systems with multi T-Flops processing power. This paper describes the main features of this new development.
As a test of quenched lattice twisted mass QCD, we compute the non-perturbatively O($a$) improved pseudoscalar and vector meson masses and the pseudoscalar decay constant down to $M_{\rm PS}/M_{\rm V} = 0.467(13)$ at $β=6$ in large volume. We check the absence of exceptional configurations and -- by further data at $β=6.2$ -- the size of scaling violations. The CPU time cost for reaching a given accuracy is close to that with ordinary Wilson quarks at $M_{\rm PS}/M_{\rm V} \simeq 0.6$ and grows smoothly as $M_{\rm PS}/M_{\rm V}$ decreases.
This paper presents the status of the APEmille project, which is essentially completed, as far as machine development and construction is concerned. Several large installations of APEmille are in use for physics production runs leading to many new results presented at this conference. This paper briefly summarizes the APEmille architecture, reviews the status of the installations and presents some performance figures for physics codes.
We present a scaling study of lattice QCD with O(a) improved Wilson fermions and a chirally twisted mass term. In order to get precise results with a moderate computational effort, we have considered a system of physical size of 0.75^3 * 1.5 fm^4 with Schroedinger functional boundary conditions in the quenched approximation. Looking at meson observables in the pseudoscalar and vector channels, we find that O(a) improvement is effective and residual cutoff effects are fairly small.
The clover coefficient $\csw$ is computed at one loop order of perturbation theory for improved gluon actions including six-link loops. The O(a) improvement coefficients for the dimension three isovector composite operators bilinear in the quark fields are also calculated.
We show that in practical simulations of lattice QCD with two dynamical light fermion species the PHMC algorithm samples configuration space differently from the commonly used HMC algorithm.
We compare the performance of the PHMC algorithm with the one of the HMC algorithm in practical simulations of lattice QCD. We show that the PHMC algorithm can lead to an acceleration of numerical simulations. It is demonstrated that the PHMC algorithm generates configurations carrying small isolated eigenvalues of the lattice Dirac operator and hence leads to a sampling of configuration space that is different from that of the HMC algorithm.
We give a detailed description of the so-called Polynomial Hybrid Monte Carlo (PHMC) algorithm. The effects of the correction factor, which is introduced to render the algorithm exact, are discussed, stressing their relevance for the statistical fluctuations and (almost) zero mode contributions to physical observables. We also investigate rounding-error effects and propose several ways to reduce memory requirements.
The numerical construction of polynomials in the product representation (as used for instance in variants of the multiboson technique) can become problematic if rounding errors induce an imprecise or even unstable evaluation of the polynomial. We give criteria to quantify the effects of these rounding errors on the computation of polynomials approximating the function $1/s$. We consider polynomials both in a real variable $s$ and in a Hermitian matrix. By investigating several ordering schemes for the monomials of these polynomials, we finally demonstrate that there exist orderings of the monomials that keep rounding errors at a tolerable level.