MINIMIZING STORAGE IN IMPLEMENTATIONS OF THE OVERLAP LATTICE-DIRAC OPERATOR

The overlap lattice-Dirac operator contains the sign function ∊(H). Recent practical implementations replace ∊(H) by a ratio of polynomials, H P_n(H²)/Q_n(H²), and require storage of 2n+2 large vectors. Here I show that one can use only four large vectors at the cost of executing the core conjugate algorithm twice. The slow-down might be less than by a factor of 2, depending on the architecture of the computer one uses.