Minimizing convex quadratics with variable precision conjugate gradients

International audienceWe investigate the method of conjugate gradients, exploiting inac-curate matrix-vector products, for the solution of convex quadratic op-timization problems. Theoretical performance bounds are derived, andthe necessary quantities occurring in the theoretical bounds estimated,leading to a practical algorithm. Numerical experiments suggest thatthis approach has significant potential, including in the steadily moreimportant context of multi-precision computations