ICL Research Profile

LAPACK

Overview

The Linear Algebra PACKage (LAPACK) and Scalable LAPACK (ScaLAPACK) are widely used libraries for efficiently solving dense linear algebra problems. ICL has been a major contributor to the development and maintenance of these two packages since their inception. LAPACK is sequential, relies on the BLAS library, and benefits from the multi-core BLAS library. ScaLAPACK is parallel, distributed, and relies on the BLAS, LAPACK, MPI, and BLACS libraries.

LAPACK 3.8.0, released in November 2017, includes level-3 BLAS communication-avoiding, symmetric-indefinite factorizations with Aasen’s triangular tridiagonalization using the two-stage algorithm. Since 2011, LAPACK has included LAPACKE, a native C interface for LAPACK developed in collaboration with Intel, which provides NAN check and automatic workspace allocation. ScaLAPACK 2.0.0, which includes the multiple relatively robust representations (MRRR) algorithm and new nonsymmetric eigenvalue problem routines, was released in November 2011. Two additional ScaLAPACK versions (2.0.1 and 2.0.2) were released in 2012 for minor bug fixes.

Find out more at http://www.netlib.org/lapack/

In Collaboration With

  1. University of California Berkeley
  2. University of Colorado Denver

Sponsored by

  1. Defense Advanced Research Projects Agency
  2. National Science Foundation
  3. The United States Department of Energy