English

The masses of the light quarks play the role of fundamental parameters of Quantum Chromodynamics (QCD), the theory of the strong interaction. Due to confinement, the phenomenon that free quarks cannot be observed, it is impossible to determine quarkmasses directly from experiments. Instead, one has to determine the dependence of certain quantities on the quarkmasses. Then the physical quarkmasses are given by comparison with physical results. The only known access to the absolute values of light quarkmasses is to evaluate QCD on a lattice. All other approaches either cannot determine absolute values (e.g. only ratios as with chiral perturbation theory) or have to use approximations with uncontrollable impact. We present new results on the mass of the light and strange quarks from SESAM/TχL. The results were obtained on lattice-volumes of 16³ x 32 and 24³ x 40 points, which allow to investigate finite-size effects. Since the SESAM/TχL ensembles with β=5.6 now are complemented by configurations with β=5.5, moreover, we are able to attempt the first continuum extrapolation of the quark masses with standard Wilson fermions. In the second part a more technical aspect of lattice QCD simulations is treated. The main computational task in full QCD simulations is the solution of large sparse linear equations on parallel computers. To improve this job, besides the choice of the solver, the optimization of the preconditioning scheme plays a major role. Here we present results obtained by the ll-SSOR preconditioner applied to O (a) improved fermions. We show that the de facto standard in todays simulations, even-odd preconditioning, is outperformed by a factor of two. Furthermore we successfully employ the ll-SSOR preconditioning scheme to hypercube fermions, a recent approach inspired by perfect actions.

Deutsch