Supercomputer Environment for Recursive Matrix Algorithms

Abstract

A new runtime environment for the execution of recursive matrix algorithms on a supercomputer with distributed memory is proposed. It is designed both for dense and sparse matrices. The environment ensures decentralized control of the computation process. As an example of a block recursive algorithm, the Cholesky factorization of a symmetric positive definite matrix in the form of a block dichotomous algorithm is described. The results of experiments with different numbers of cores are presented, demonstrating good scalability of the proposed solution.