Evaluation Of Large-Scale Optimization Problems On Vector And Parallel Architectures

. We examine the importance of problem formulation for the solution of large-scale optimization problems on high-performance architectures. We use limited memory variable metric methods to illustrate performance issues. We show that the performance of these algorithms is drastically affected by application implementation. Model applications are drawn from the MINPACK-2 test problem collection, with numerical results from a super-scalar architecture (IBM RS6000/370), a vector architecture (CRAY-2), and a massively parallel architecture (Intel DELTA). Key words. optimization, large-scale, limited memory, variable metric, performance evaluation, vector architecture, parallel architecture. AMS subject classifications. 65Y05, 65Y20, 65K05, 65K10, 90C06, 90C30 1. Introduction. Our aim is to explore performance issues associated with the solution of large-scale optimization problems on high-performance architectures. The solution of these problems, where the number of variables ranges betwe..