Mathematical Libraries
Table of Contents
- Introduction
- Comprehensive mathematical libraries
- Differential Equation Solvers
- Fast Fourier transforms
- Linear algebra
- Mesh partitioning
- Statistics
Introduction
Brief notes are given below on installed mathematical libraries and applications, including numerical libraries for linear algebra, Fast Fourier transforms and statistical analysis. See the main software page for a comparison of which versions are installed on the various WestGrid systems.
MATLAB, a high-level programming environment for scientific computation, which contains many useful linear algebra and other mathematical rouintes, is described on a separate page.
Comprehensive mathematical libraries
ESSL - IBM Engineering and Scientific Subroutine Library.
This numerical library contains linear algebra routines (BLAS, most of LAPACK, eigensystem analysis. Fourier transforms, convolutions and correlations, sorting, interpolation, quadrature and random number generation. Please note that routines in ESSL have the same names as LAPACK routines but a different calling sequence. Do not link LAPACK calls to the ESSL libraries unless you have checked to make sure they are equivalent. Some of the routines can use multiple threads in OpenMP programs. There is a also a Parallel ESSL containing ScaLAPACK compatible routines, Fourier transforms and random number generation for use in MPI programs. See the IBM documentation for more information.
GSL - GNU Scientific Library.
As described on the GSL home page, it is a numerical library for C and C++ programmers including numerical integration, linear algebra, minimization, special functions and other mathematical routines.
MKL - Intel Math Kernel Library.
This numerical library contains BLAS (including a C interface), LAPACK, a sparse system solver, random number generators and vector versions of common mathematical functions. See the Intel documentation for more information.
It is installed on glacier in /global/software/intel/mkl61, where one can find examples and additional documentation.
SCSL - Scientific Computing Software Library.
This package for SGI IRIX computers contains BLAS and LAPACK linear algebra routines, sparse system solvers, and signal processing (FFT and convolution) routines. See the SGI documentation for more information.
Differential Equation Solvers
PETSc - A toolkit for solution of differential equations with parallel solvers.
See the PETSc home page and some notes on PETSc on WestGrid.
Fast Fourier transforms
FFTW - A widely-used FFT implementation.
See the FFTW home page.
Linear algebra
BLAS - Basic Linear Algebra Subprograms
BLAS are fundamental routines for vector and matrix operations.
See the Netlib BLAS page for background information and references. Optimized BLAS libraries are available for all WestGrid systems as part of other packages. Many users will call higher level routines, such as those in LAPACK, rather than using the BLAS routines directly. BLAS are Fortran routines, but, C interfaces are available for some systems.
LAPACK - Linear algebra subroutine package.
See the Netlib LAPACK page for a user guide as well as man pages on the individual WestGrid machines. Also, see the Comprehensive mathematical libraries section above.
PIM - Parallel Iterative Methods
PIM is a Fortran 77 package for solving systems of linear equations. See the PIM home page for documentation.
ScaLAPACK - Scalable LAPACK
ScaLAPACK is a parallelized subset of the LAPACK linear algebra package. See the Netlib ScaLAPACK page for documentation and tutorials.
On nexus.westgrid.ca, see the README and examples in /usr/global/scalapack64 .
WSMP - Watson Sparse Matrix Package.
This package contains libraries for the direct solution of sparse systems of linear equations (for both serial and parallel environments). See the IBM WSMP page or the WSMP home page for more information.
Mesh partitioning
METIS
METIS is a library and a family of programs for partitioning meshes and graphs. See the METIS home page.
Statistics
R
R is a software environment and language for statistical data analysis. See the R home page.
Updated 2009-11-20.