In this paper we propose a coordinate descent algorithm for solving high dimensional risk parity problems. We show that this algorithm converges and is very fast even with large covariance matrices (n500). Comparison with existing algorithms also shows that it is one of the most efficient...