PROCESSORS

The Numerical Algorithms Group, a worldwide organization dedicated to developing quality mathematical, statistical and 3D visualization software components, is announcing a major new release of the NAG SMP Library at Supercomputing 2003. The significant additions to this Release 3 of the NAG SMP Library further extend the abilities of developers and programmers to make optimal use of the processing power and shared memory parallelism of their Symmetric Multi-Processor (SMP) systems. The NAG Fortran SMP Library, with over 1300 routines, now contains over 250 routines that have been optimized or enhanced for use on SMP systems. This includes 25 additional optimized routines added at Release 3. These additions include: • ARPACK additions for the iterative computation of large sparse eigensolutions (Sorensen et al); • SuperLU, specifically large sparse direct solver routines (Demmel et al, Berkeley); • LAPACK drivers and new computational routines from LAPACK 3; • Complex sparse iterative solvers; • Solvers for banded linear systems of equations; • Bandwidth reduction routines; • Pre-conditioners. The major algorithmic content of Release 3 NAG SMP Library also appears in the latest releases of the NAG Fortran and C Libraries for single processors, but with identical calling sequences. This enables developers to use the same line of code to call to a routine that performs the same function, in either the serial library or the Fortran SMP Library. As with earlier releases, Release 3 of the NAG SMP library achieves significant performance gains over the hardware vendors' SMP implementations and has proven scalability up to 64 processors.