SuperLU

Sparse Gaussian Elimination on High Performance Computers

Jim Demmel

Xiaoye Li

SuperLU for Sequential Machines

This version of SuperLU is optimized for high performance on superscalar uniprocessors with caches.
  • Downloadable software and documentation
  • A Supernodal Approach to Sparse Partial Pivoting, by J. Demmel, S. Eisenstat, J. Gilbert, X. Li, and J. Liu. to appear in SIAM J. Mat. Anal. Appl.
  • Slides from a talk given by Xiaoye Li at Stanford 10/97 (covers both SuperLU and SuperLU_MT)

    • SuperLU_MT for Shared Memory Parallel Processors (SMPs)

    This version of SuperLU is optimized for high performance on SMPs.
  • Downloadable software and documentation
  • An Asynchronous Parallel Supernodal Algorithm for Sparse Gaussian Elimination, by J. Demmel, J. Gilbert, and X. Li. to appear in SIAM J. Mat. Anal. Appl.
  • Slides from a talk given by Xiaoye Li at Stanford 10/97 (covers both SuperLU and SuperLU_MT)

    • SuperLU_Dist for Distributed Memory Parallel Processors

    This version of SuperLU is optimized for high performance on distributed memory processors. It will be released in early 1999.
  • Making Sparse Gaussian Elimination Scalable by Static Pivoting, by J. Demmel and X. Li. SuperComputing 98. (includes the full paper, slides, and a table of numerical results)