In order to compute physical quantities in lattice quantum chromodynamics huge systems of linear equations have to be solved. The availability of eecient parallel Krylov subspace solvers plays a vital role in the solution of these systems. We present a detailed analysis of the performance of the stabilized biconjugate gradient (BiCGStab) algorithm with… (More)
The calculation of physical quantities by lattice QCD simulations requires in some important cases the determination of the inverse of a very large matrix. In this article we describe how stochastic estimator methods can be applied to this problem, and how such techniques can be efficiently implemented on parallel computers.
Large scale simulations of Quantum Chromodynamics are extremely CPU time demanding. Today, simulations in a heterogeneous SIMD-MIMD environment can be an interesting step towards simulations on Teracomputers, as already available for Japanese and US groups.
The cost for stochastic sampling of QCD vacuum conngurations outweighs by far the costs of the remaining computational tasks in Lattice QCD, due to the nonlocal forces arising from the dynamics of fermion loops in the vacuum uctuations. The evaluation of quality and hence eeciency of sampling algorithms is largely determined by the assessment of their… (More)