superlu-dist: Sparse Linear system solver (distributed mem/MPI)

Contents

  1. Overview of package
  2. Overview of package
    1. General usage
  3. Availability of package by cluster

Overview of package

General information about package
Package: superlu-dist
Description: Sparse Linear system solver (distributed mem/MPI)
For more information: https://github.com/xiaoyeli/superlu
Categories:
License: OpenSource (LLNL License)

General usage information

SuperLU contains a set of subroutines to solve a sparse linear system A*X=B. It uses Gaussian elimination with partial pivoting (GEPP).

This is the distributed memory parallelism implementation' of the library. I.e., it uses MPI to allow processes to be spread over cores of multiple nodes.

Please see superlu for a sequential version.

Please see superlu-mt for a multithreaded/shared memory parallelism version (i.e. can use multiple cores on a single node).

The following environmental variables have been defined:

  • \$SUPERLUDIST_ROOT has been set to the root of the superlu-dist installation
  • \$SUPERLUDIST_LIBDIR points to the directory containing the libraries
  • \$SUPERLUDIST_INCDIR points to the directory containing the header files

You will probably wish to use these by adding the following flags to your compilation command (e.g. to CFLAGS in your Makefile):

  • -I\$SUPERLUDIST_INCDIR
and the following flags to your link command (e.g. LDFLAGS in your Makefile):
  • -L\$SUPERLUDIST_LIBDIR -Wl,-rpath,\$SUPERLUDIST_LIBDIR

Available versions of the package superlu-dist, by cluster

This section lists the available versions of the package superlu-diston the different clusters.

Available versions of superlu-dist on the Deepthought2 cluster (RHEL8)

Available versions of superlu-dist on the Deepthought2 cluster (RHEL8)
Version Module tags CPU(s) optimized for GPU ready?
5.3.0 superlu-dist/5.3.0 ivybridge Y
6.3.0 superlu-dist/6.3.0 ivybridge, x86_64 Y

Available versions of superlu-dist on the Juggernaut cluster

Available versions of superlu-dist on the Juggernaut cluster
Version Module tags CPU(s) optimized for GPU ready?
5.3.0 superlu-dist/5.3.0 skylake_avx512, x86_64, zen Y
6.3.0 superlu-dist/6.3.0 skylake_avx512, x86_64, zen Y