Efficient vector and parallel manipulation of tensor products

  • Authors:
  • Paul E. Buis;Wayne R. Dyksen

  • Affiliations:
  • Ball State Univ., Muncie, IN;Ball State Univ., Muncie, IN

  • Venue:
  • ACM Transactions on Mathematical Software (TOMS)
  • Year:
  • 1996

Quantified Score

Hi-index 0.00

Visualization

Abstract

We present efficient vector and parallel methods formanipulating tensor products of matrices. We consider bothcomputing the matrix-vector product (A1ļÄAK)x and solving the systemof linear equations (A1ļÄAK)x=b. The methods describedare independent of K. We accompany this article with acompanion algorithm which describes an implementation of a completeset of tensor product routines based on LAPACK and the Level 2 and3 Basic Linear Algebra Subprograms (BLAS) which providevectorization and parallelization.