Anatomy of a Parallel Out-of-Core Dense Linear Solver

In this paper, we describe the design and implementation of the Platform Independent Parallel Solver (PIPSolver) package for the out-ofcore (OOC) solution of complex dense linear systems. Our approach is unique in that it allows essentially all of RAM to be filled with the current portion of the matrix (slab) to be updated and factored, thereby greatly improving the computation to I/O ratio over previous approaches. Experiences and performance are reported for the Cray T3D system. INTRODUCTION Scalable, portable linear algebra libraries for distributed memory MIMD architectures that are currently being investigated put a limit on the size of the problem that can be solved, dictated by the aggregate RAM of the machine. A good example of this is the ScaLAPACK library for dense linear algebra [3]. This paper is concerned with the investigation of techniques for development of a dense linear solver package that is scalable both in machine and problem size. A number of parallel out-of-core..