1978
DOI: 10.1145/355780.355783
|View full text |Cite
|
Sign up to set email alerts
|

The PORT Mathematical Subroutine Library

Abstract: The development at Bell Laboratorms of PORT, a hbrary of portable Fortran programs for numemcal computation, is discussed. Portab]hty is achmved by careful language specification, together with the key techmque of spemfymg computer classes by means of predefined machine constants The library is built around an automatm error-handling facility and a dynamm storage allocatmn scheme, both of which are implemented portably These, together with the modular structure of the library, lead to slmphfied calhng sequence… Show more

Help me understand this report

Search citation statements

Order By: Relevance

Paper Sections

Select...
2
1
1
1

Citation Types

0
83
0
1

Year Published

1978
1978
2014
2014

Publication Types

Select...
5
4

Relationship

0
9

Authors

Journals

citations
Cited by 139 publications
(89 citation statements)
references
References 25 publications
0
83
0
1
Order By: Relevance
“…CALGO has seen the publication of algorithms that have had a major impact on the way in which we perform numerical computation; for example, the various families of Basic Linear Algebra Subroutines [38,13,12] and the use of a standard function for accessing machine dependent parameters [19].…”
Section: Resultsmentioning
confidence: 99%
“…CALGO has seen the publication of algorithms that have had a major impact on the way in which we perform numerical computation; for example, the various families of Basic Linear Algebra Subroutines [38,13,12] and the use of a standard function for accessing machine dependent parameters [19].…”
Section: Resultsmentioning
confidence: 99%
“…Values for these constants are given for many computers in the in-line documentation. SPMPAR, released by Argonne National Laboratory, is an adaptation of the Bell Laboratories function RlMACH [2].…”
Section: Machine-dependent Constantsmentioning
confidence: 99%
“…The coefficients are determined by a Galerkin method [13] g(xm, Vm , u{x m , yTO )) = 0, m in D. (17) Equation (16) …”
Section: =1mentioning
confidence: 99%
“…To calculate v, replace u by u + v in eqs (17) and (18) Substituting for u and v from eqs (15) and (19) Using this procedure, the solution of eqs (1) and (7) is started by first obtaining a trial solution for the linearized form of eq (1), using the solution of eq (7) obtained from the one-dimensional charge-sheet model with the total charge as a boundary condition. The Jacobian of g required for the solution of eq (20) contains the Jacobian of eq (1).…”
Section: =1mentioning
confidence: 99%