Linear algebra software in IBM and CRAY computers
This report is concerned with the performance (accuracy and runtimes) of linear algebra software from IMSL (Edition 10) and NAG (Mark 12) and of some other programs. All tests have been run on IBM 3081 (in double precision) and CRAY X-MP/22 (in single precision). The software is related to matrix multiplication, linear systems with symmetric positive definite matrix (algorithms based on the Cholesky decomposition) and with real square matrix (LU-decomposition), complex linear systems, real (symmetric) eigenvalue problems. The report shows that, for the test problems (special attention has been paid to large-order test matrices), the library routines do not differ significantly in the accuracy of the computed results. But differences can be observed for the CPU-time spent to solve some particular problems.
- Basic Linear Algebra Subprograms for Fortran Usage
- Corrigenda: “An Extended Set of FORTRAN Basic Linear Algebra Subprograms”
- scientific article; zbMATH DE number 3748409 (Why is no real title available?)
- scientific article; zbMATH DE number 3311329 (Why is no real title available?)
- scientific article; zbMATH DE number 3349178 (Why is no real title available?)
- Implementing Linear Algebra Algorithms for Dense Matrices on a Vector Pipeline Machine
- Linear algebra software on a vector computer
- Matrix eigensystem routines - EISPACK guide. 2nd ed
- Solving Linear Algebraic Equations on an MIMD Computer
- Unrolling loops in fortran
This page was built for publication: Linear algebra software in IBM and CRAY computers
Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q1822892)