Abstract
Parallel computing systems are now very common; essentially all new systems with any pretence at providing scientific computing facilities contain a number of processors which can cooperate to run a single program. “Highly parallel” computing systems are those for which the number of processors is so large that the treatment of parallelism must be systematic rather than ad hoc; it seems likely that such systems will rapidly become the norm, and it is therefore necessary to provide a range of utility software for them. We consider here the provision of numerical library facilities for parallel systems. The emphasis is on work being carried out under the EEC-funded Esprit 1 programme on the production of a numerical library for transputer arrays; topics covered include:
-
Library structure, and the programming interface presented to the user
-
Algorithm design criteria
-
Achievable efficiencies
-
Progress to date.
The work involved in the production of a comprehensive library is very large; we therefore consider also:
-
Portability between architectures
-
Portability between languages.
This is a preview of subscription content, log in via an institution to check access.
Access this chapter
Tax calculation will be finalised at checkout
Purchases are for personal use only
Preview
Unable to display preview. Download preview PDF.
References
Brown, N.G. and Delves, L.M. (1988), A Dynamic Library Loading Mechanism for Transputer arrays. P1085 Working Paper, Liverpool University, March 1988.
Brown, N.G. and Delves, L.M. (1988a), A Dynamic Loading Mechanism for Transputer arrays, Esprit P1085 Working Paper, Liverpool June 1988.
Delves, L.M. (1987) An Error Mechanism for Ada Numerical Libraries, in: Scientific Ada, ed. B. Ford, J. Kok and M. Rogers, Ada Companion Series, Cambridge University Press, 1987.
Delves, L.M. and Brown, N.G. (1987), An Error Mechanism for the occam Library, P1085 working paper, University of Liverpool 1987.
Delves, L.M. and Brown, N.G. (1988), A Numerical Library for Transputer Arrays, Invited paper presented at IFIP WG2.5 Working Conference on “Aspects of Computation on Asynchronous Parallel Processors”, August 1988, Stanford, USA; proceedings to be published.
Howard, G. (1987), Gauss Elimination on Transputer Arrays, P1085 Working Paper, September 1987; to be published.
Inmos (1987), The Occam 2 Reference Manual, INMOS Ltd.
May, D. (1986), Occam 2 Language Definition, INMOS Ltd.
May, D. (1986a), The Transputer, Unicom Seminar on Major Advances in Parallel Processing, London 1986.
Lawson, C.L., Hanson, K.J., Kincaid, D.R. and Krogh, F.T. (1979), Basic Linear Algebra Subroutines for Fortran Usage, ACM Trans. Math. Software 5, pp. 308–323.
Dongarra, J.J., du Croz, J., Duff, I., and Hammarling, S. (1987), Technical Memo 88, Argonne National Laboratory, Argonne, 111.
Author information
Authors and Affiliations
Editor information
Editors and Affiliations
Rights and permissions
Copyright information
© 1990 Chapman and Hall
About this paper
Cite this paper
Delves, L.M. (1990). Scientific library software for highly parallel computers. In: Mason, J.C., Cox, M.G. (eds) Scientific Software Systems. Springer, Dordrecht. https://doi.org/10.1007/978-94-009-0841-3_6
Download citation
DOI: https://doi.org/10.1007/978-94-009-0841-3_6
Publisher Name: Springer, Dordrecht
Print ISBN: 978-94-010-6865-9
Online ISBN: 978-94-009-0841-3
eBook Packages: Springer Book Archive