• QUICK LINKS
  • answers.stanford.edu
  • apple.stanford.edu
  • computing.stanford.edu
  • courses.stanford.edu
  • datarates.stanford.edu
  • dell.stanford.edu
  • departmentphone.stanford.edu
  • email.stanford.edu
  • ess.stanford.edu
  • helpsu.stanford.edu
  • myitserviceshelp.stanford.edu
  • orderithelp.stanford.edu
  • software.stanford.edu
  • studentphone.stanford.edu
  • techtraining.stanford.edu
  • tools.stanford.edu
  • unixcomputing.stanford.edu
  • vpn.stanford.edu
  • web.stanford.edu
  • windows.stanford.edu
• SERVICES
  • IT Services Only
  • All Campus Providers
    → computing.stanford.edu
• GETTING HELP
  • Stanford Answers
  • HelpSU: IT Help Desk
  • Training
• PROJECTS
• ABOUT US
  • What We Do
  • Organization Chart
  • Strategic Plan
  • Jobs
  • Contact info
  • Service & System Metrics
• INTERNAL
• SEARCH

  • Search IT Services:
     

• Pubsw Software
• Index of Leland Software
• LAPACK 3.0

LAPACK 3.0

Summary

Description: FORTRAN numerical linear algebra library
Source: http://www.netlib.org/lapack/

Support

This package is available for all platforms. This package is unsupported and provided on an as-is basis.

Description

LAPACK provides routines for solving systems of simultaneous linear equations, least-squares solutions of linear systems of equations, eigenvalue problems, and singular value problems. The associated matrix factorizations (LU, Cholesky, QR, SVD, Schur, generalized Schur) are also provided, as are related computations such as reordering of the Schur factorizations and estimating condition numbers. Dense and banded matrices are handled, but not general sparse matrices. In all areas, similar functionality is provided for real and complex matrices, in both single and double precision.

LAPACK started as a reimplementation of the widely-used EISPACK and LINPACK routines to run more efficiently on shared-memory vector and parallel processors. As much of the computation as possible is performed by the BLAS library, which may be a highly optimized implementation provided with the operating system.

Usage

lapack.a is installed as /usr/pubsw/lib/liblapack.a and tmglib.a is installed as /usr/pubsw/lib/libtmg.a. With most compilers, these libraries can be linked against by using:

    -L /usr/pubsw/lib -llapack (or -ltmg)

for the link flags. Whenever possible, these routines are built using the vendor Fortran compiler and the system BLAS libraries for speed. You may also have to link against an appropriate BLAS library when linking against LAPACK, since LAPACK uses BLAS.

Look at the makefile fragments in /usr/pubsw/doc/Data/lapack/makefiles to see the compiler, compiler flags, and BLAS library used for each platform.

On Solaris, it's best to either link with a Fortran compiler if possible using the -xlic_lib=sunperf flag, since that will pick up the right libraries. When compiling with C, try:

    -L/usr/pubsw/lib -L/opt/SUNWspro/lib -llapack -lsunperf -lfsu -lF77
      -lsunmath -lm

This is known to work with the Sun compiler (cc).

Files

  Libraries: libblas.a liblapack.a libtmg.a

Accuracy

This information reflects the version of this package that is installed in Stanford's site-wide pubsw software tree. Newer versions of the package may be available from the package source. The contents of this page, including the source URL if applicable, are generally only reviewed when the package is upgraded and therefore may not accurately reflect the current version of the package.