NAG Parallel Library 3.0 with new features

This latest release of the NAG Parallel Library contains 183 parallelised algorithms of which 95 are new. In this release NAG now provides a further 10 routines which offer extra functionality for generating random numbers in parallel, assisting large scale simulation. This enhanced functionality includes the generation of sequences of independent pseudo random numbers, e.g. Monte Carlo simulations used by finance experts.

For Engineering an extensive range of Dense Linear Algebra solvers have been included, many of which are based on ScaLAPACK; these routines can also be applied to a broad range of scientific problems. 5 newly parallelised routines for SVD problems have been added. These routines are commonly used by engineers, especially those looking to solve problems in modelling and structural engineering. Also 26 Sparse Linear Algebra routines, newly parallelised for the Distributed Memory machines will benefit engineers with sparse matrix problems, particularly those working in design and manufacturing for example in the Automotive and Aerospace industries and other users with large scale simulation and PDE problems.

The library has been designed for users with problems that are large enough to make efficient use of the increased processing power and memory capacity of multiple processors. Anyone using a distributed memory, hybrid machine or SMP for example, people with very large datasets or carrying out simulation. Users in the areas of weather forecasting, large-scale CAD, automotive/aerospace design, energy providers, oil exploration will also benefit from this new release. As always users can rest assured that every algorithm has been fully tested, validated and verified. Several routines in this latest release have been developed and tested in industrial applications by members of the highly successful PINEAPL consortium, including Piaggio, British Aerospace, Thomson LCR, IBM SEMEA and the Danish Hydraulic Institute.