In this paper we investigate the benefits of using Z(2)xZ(2) single timeslice stochastic sources for the calculation of light quark physics on the lattice. Meson 2-point correlators measured using sources stochastic in only spin and those stochastic
in both spin and colour indices are compared to point source correlators on the unit gauge and on a 16^3 x 32 Domain Wall QCD ensemble. It is found that the use of stochastic sources gives a considerable improvement in statistics for the same computational cost. The neutral kaon mixing matrix element B_K is also calculated on this ensemble with stochastic sources, but we conclude that the stochastic method offers no significant advantage over the traditional gauge-fixed wall source approach which already offers an exact volume average. We also discuss the application to semileptonic form factors in conjunction with partially twisted boundary conditions.
We present the first results for the Kl3 form factor from simulations with 2+1 flavours of dynamical domain wall quarks. Combining our result, namely f_+(0)=0.964(5), with the latest experimental results for Kl3 decays leads to |V_{us}|=0.2249(14), r
educing the uncertaintity in this important parameter. For the O(p^6) term in the chiral expansion we obtain Delta f=-0.013(5).
We compute the electromagnetic form factor of a pion with mass m_pi=330MeV at low values of Q^2equiv -q^2, where q is the momentum transfer. The computations are performed in a lattice simulation using an ensemble of the RBC/UKQCD collaborations gaug
e configurations with Domain Wall Fermions and the Iwasaki gauge action with an inverse lattice spacing of 1.73(3)GeV. In order to be able to reach low momentum transfers we use partially twisted boundary conditions using the techniques we have developed and tested earlier. For the pion of mass 330MeV we find a charge radius given by <r_pi^2>_{330MeV}=0.354(31)fm^2 which, using NLO SU(2) chiral perturbation theory, extrapolates to a value of <r_pi^2>=0.418(31)fm^2 for a physical pion, in agreement with the experimentally determined result. We confirm that there is a significant reduction in computational cost when using propagators computed from a single time-slice stochastic source compared to using those with a point source; for m_pi=330MeV and volume (2.74fm)^3 we find the reduction is approximately a factor of 12.
