
In this second report on our recent numerical simulations of twoflavour 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 DDHMC parameters, the issue of stability, autocorrelations and the
statistical error analysis. Extensive data tables are included as well as a
short discussion of the quarkmass dependence in partially quenched QCD,
supplementing the physics analysis that was presented in the first paper in
this series.

Recent conceptual, algorithmic and technical advances allow numerical
simulations of lattice QCD with Wilson quarks to be performed at significantly
smaller quark masses than was possible before. Here we report on simulations of
twoflavour QCD at seaquark masses from slightly above to approximately 1/4 of
the strangequark mass, on lattices with up to 64x32^3 points and spacings from
0.05 to 0.08 fm. Physical seaquark effects are clearly seen on these lattices,
while the lattice effects appear to be quite small, even without O(a)
improvement. A striking result is that the dependence of the pion mass on the
seaquark mass is accurately described by leadingorder chiral perturbation
theory up to meson masses of about 500 MeV.

We study the spectral gap of the WilsonDirac operator in twoflavour
lattice QCD as a function of the lattice spacing $a$, the spacetime volume $V$
and the currentquark mass $m$. It turns out that the median of the probability
distribution of the gap scales proportionally to $m$ and that its width is
practically equal to $a/\sqrt{V}$. In particular, numerical simulations are
safe from accidental zero modes in the largevolume regime of QCD.

In lattice QCD it is possible, in principle, to determine the parameters in
the effective chiral lagrangian (including weak interaction couplings) by
performing numerical simulations in the $\epsilon$regime, i.e. at quark
masses where the physical extent of the lattice is much smaller than the
Compton wave length of the pion. The use of a formulation of the lattice theory
that preserves chiral symmetry is attractive in this context, but the numerical
implementation of any such approach requires special care in this kinematical
situation due to the presence of some very low eigenvalues of the Dirac
operator. We discuss a set of techniques (lowmode preconditioning and
adaptedprecision algorithms in particular) that make such computations
numerically safe and more efficient by a large factor.

The gauge covariant lattice Dirac operator D which has recently been proposed
by Neuberger satisfies the GinspargWilson relation and thus preserves chiral
symmetry. The operator also avoids a doubling of fermion species, but its
locality properties are not obvious. We now prove that D is local (with
exponentially decaying tails) if the gauge field is sufficiently smooth at the
scale of the cutoff. Further analytic and numerical studies moreover suggest
that the locality of the operator is in fact guaranteed under far more general
conditions.

The past few years have seen many interesting theoretical developments in
lattice QCD. This talk (which is intended for nonexperts) focuses on the
problem of nonperturbative renormalization and the question of how precisely
the continuum limit is reached. Progress in these areas is crucial in order to
be able to compute quantities of phenomenological interest, such as the hadron
spectrum, the running quark masses and weak transition matrix elements, with
controlled systematic errors.

It is shown how onshell O(a) improvement can be implemented
nonperturbatively in lattice QCD with Wilson quarks. Improvement conditions
are obtained by requiring the PCAC relation to hold exactly in certain matrix
elements. These are derived from the QCD Schr\"odinger functional which enables
us to simulate directly at vanishing quark masses. In the quenched
approximation and for bare couplings in the range $0\leq g_0\leq 1$, we
determine the improved action, the improved axial current, the additive
renormalization of the quark mass and the isospin current normalization
constants Z_A and Z_V.

The conservation of the isovector axial current in lattice QCD with massless
Wilson quarks is studied to oneloop order of perturbation theory. Following a
strategy described in a previous publication, the O($a$) counterterm required
for onshell improvement of the current is computed. We also confirm an earlier
result of Wohlert for the coefficient in front of the SW term in the improved
action.

A previously introduced multiboson technique for the simulation of QCD with
dynamical quarks is described and some results of first test runs on a
$6^3\times12$ lattice with Wilson quarks and gauge group SU(2) are reported.