Basic row-column operations with orthogonally linked sparse matrices for use with FORTRAN
An orthogonally linked data structure is used to represent sparse rectangular matrices. This representation requires four type INTEGER locations and one type REAL location per nonzero entry in the matrix. Operations are presented that perform many of the basic computations required in numerical linear algebra. The operations are available in portable FORTRAN, with the exception of the necessarily operating-system-sensitive ''get'' and ''put'' subprogram for the individual records. The package of subprograms manages out-of-memory portions of the matrices as well as the insertion, modification, and deletion of entries. By use of the package, many of the familiar processes of numerical linear algebra, such as solving sparse linear algebraic equations, can be written with a relatively few calls to these subprograms. 2 figures, 3 tables.
- Research Organization:
- Sandia Labs., Albuquerque, N.Mex. (USA)
- DOE Contract Number:
- EY-76-C-04-0789
- OSTI ID:
- 7209490
- Report Number(s):
- SAND-77-1091
- Country of Publication:
- United States
- Language:
- English
Similar Records
Basic linear algebra subprograms for FORTRAN usage. [BLAS, in FORTRAN and assembly language for IBM 360/67, CDC 6600 and 7600, and Univac 1108]
Threshold partitioning of sparse matrices and applications to Markov chains