Parallel Tensor Methods for Nonlinear Equations and Nonlinear Least Squares

We describe the design and computational performance of parallel row-orientedtensor algorithms for the solution of dense systems of nonlinear equations and nonlinearleast squares problems on a distributed-memory MIMD multiprocessor. Tensormethods are general purpose methods that base each iteration upon a quadratic modelof the nonlinear function, rather than the standard linear model, where the secondorder term is selected so that the model is hardly more expensive to form, store, orsolve ...