Abstract
The requirements for supercomputing in technical sciences in an industrial R&D environment or in a versatile job profile university environment are specified: 100 GFLOPS sustained performance, 64 Gwords main memory, flexible data transfer operations (compress, expand, merge, gather, scatter), portable Fortran 8x, fastest scalar speed. Then the reasons are discussed why the”usual” trend to parallelism via MM) (message passing systems, shared memory systems, hybrid systems) fail to meet the requirements of the users. The proposition of a Continuous Pipe Vector Computer (CPVC) serves to explain in the form of 10 notes the ideas how parallelism should be organized that it is completely transparent to the user. The proposed CPVC minimizes the lost cycles of a supercomputer so that one gets close to the theoretical peak performance by the most user-friendly architecture.
With kind permission of Computing Center, University of Karlsruhe, where this Contribution has been published as Internal Report Nr. 35/89, January 1989, together with an appendix’Epilog’
Access this chapter
Tax calculation will be finalised at checkout
Purchases are for personal use only
Preview
Unable to display preview. Download preview PDF.
Literatur
W. Schönauer, Scientific Computing on Vector Computers, North-Holland, Amsterdam, New York, 1987
Proceedings of the”2nd International SUPRENUM Colloquium 1987”, Bonn, Sept. 30 to Oct. 2, 1987, to appear as special issue of “Parallel Computing”
I.S. Duff, A survey of Supercomputing in Europe, to appear in [2]
W. Schönauer, W. Gentzsch, The Efficient Use of Vector Computers with Emphasis to Computational Fluid Dynamics, Vieweg, Braunschweig/Wiesbaden, 1986
W. Schönauer, E. Schnepf, FIDISOL, a”black box” solver for partial differential equations, to appear in Parallel Computing
W. Schönauer, E. Schnepf, H. Müller, The FIDISOL Program Package, Interner Bericht Nr. 27/85 des Rechenzentrums der Universität Karlsruhe, 1985. This internal report is the documentation for the customers of FIDISOL.
W. Schönauer, E. Schnepf, H. Müller, Designing PDE software for vector computers as a “data flow Algorithm”, Computer Physics Communications 37 (1985), pp. 233–237 and
W. Schönauer, E. Schnepf, H. Müller, Designing PDE software for vector computers as a “data flow Algorithm”, I.S. Duff, J.K. Reid (Eds), Vector and Parallel Processors in Computational Science, North-Holland, Amsterdam, New York 1985, pp. 233–237
M. Metcalf, Fortran 8x — the emerging standard, Computer Physics Communications 45 (1987), pp. 259–268
Fortran, X3J3/S8.104, June 1987, American National Standards Institute. This is the actual draft under discussion for Fortran 8x.
K. Hwang, S.M. Jacobs, E.E. Swartzlander (Eds), Proceedings of the 1986 Internat. Conf. on Parallel Processing, IEEE, Washington D.C., 1986
J.L. Gustafson, S. Hawkinson, K. Scott, The architecture of a homogeneous vector supercomputer, in [10], pp. 649–652
E. Schmidt, Rechnergiganten aus dem Baukasten, VDI nachrichten 16, 18. April 1986, p. 17
J.J. Dongarra (Ed), Experimental Parallel Computing Architectures, North-Holland, Amsterdam, New York 1987
D. Degroot (Ed), Proceedings of the 1985 Internat. Conf. on Parallel Processing, IEEE, Washington, D.C., 1985.
Proceedings of the Int. Conf. on Vector and Parallel Computing, Loen, Norway, Parallel Computing 5 (1987) pp. 1–263
Proceedings of the 1984 IBM Europe Institute course on Highly Parallel Processing, Parallel Computing 2 (1985), pp. 185–288
M.T. Heath (Ed), Hypercube Multiprocessors 1987, SIAM, Philadelphia, 1987
W. Giloi, The SUPRENUM architecture, to appear in [2]
V.P. Srini, Anarchitectural comparison of dataflow systems, Computer, vol 19, No. 3 (1986) pp. 68–88
J. Gurd, C. Kirkham, W. Böhm, The Manchester dataflow computing system, in [13], pp. 177–219
P.M. Behr, W.K. Giloi, H. Mühlenbein, SUPRENUM: The German supercomputer architecture — rationale and concepts, in [10], pp. 567–575
Alliant Computer Systems Corporation, Acton, Mass., FX/ Series Product Summary, 1985
W.D. Hillis, The Connection Machine, MIT press, Cambridge, Mass. 1985
G.F. Pfister, W.C. Brantley, D.A. George, L.S. Harvey, W.J. Kleinfelder, K.P. McAuliffe, E.A. Melton, V.A. Norton, J. Weiss, An introduction to the IBM Research Parallel Processor Prototype (RP3), in [13], pp. 123–140
G.C. Fox, Questions and unexpected answers in concurrent computation, in [13], pp. 97–121
Intel, iPSC User,s Guide, Intel, Portland, Oregon, 1985
M.C. Chen, Very-high-level parallel programming in Crystal, in [17], pp. 39–47
W. Williams, Load balancing and Hypercubes: A preliminary look, in [17], pp. 108–113
J.H. Sultz, M.C. Chen: Automated problem mapping: The Crystal runtime system, in [17], pp. 130–140
K. Schwan, W. Bo, N. Bauman, P. Sadayappan, F. Ercal, Mapping parallel applications to a hypercube, in [17], pp.141–151
D.W. Walker, G.C. Fox, A. Ho, G.R. Montry, A comparision of the performance of the Caltech Mark II hypercube and the Elxsi 6400, in [17], pp. 210–219
J.M. Francioni, J.A. Jackson, An implementation of a 2 d -section root finding method for the FPS T-series hypercube, in [17], pp. 495–500
R.M. Chamberlain, An alternative view of LU factorization with partial pivoting on a hypercube multiprocessor, in [17], pp. 569–575
V.K. Naik, S. Taasan, Performance studies of the multigrid algorithms implemented on hypercube multiprocessor systems, in [17], pp. 720–729
D.J. Kuck, E.S. Davidson, D.H. Lawrie, A.H. Sameh, Parallel supercomputing today and the CEDAR approach, in [13], pp. 1–23
A. Gottlieb, An overview of the NYU Ultracomputer project, in [13], pp. 25–95
M.J. Flynn, Some computer organizations and their effectiveness, IEEE Trans. Comput. C-21 (1972), pp. 948–960
B.L. Buzbee, Applications of MMD machines, Computer Physics Communications 37 (1985), pp. 1–5, or
B.L. Buzbee, Applications of MMD machines, LS. Duff, J.K. Reid (Eds), Vector and Parallel Processors in Computational Science, North-Holland, Amsterdam, New York, Oxford, Tokyo 1985, pp. 1–5
R.W. Hockney, (r,m1/2, s1/2)measurements on the 2-CPU CRAY X-MP, Parallel Computing, vol 2, Nr. 1, March 1985, pp.1–14
A.K. Dave, The efficient use of the CRAY X-MP multiprocessor vector computer in computational fluid dynamics, in [4], pp. 209–220
R.W. Hockney, C.R. Jesshope, Parallel Computers, Adam Hilger, Bristol, 1981
W. Myers, Getting the cycles out of a supercomputer, Computer, vol 19 (1986), pp. 89–92
A.H. Karp, Programming for Parallelism, Computer, vol 20 (1987), pp. 43–57
E. Clementi, J. Detrich, Large scale parallel computation on a loosely coupled array of processors, in [13], pp. 141–176
R.W. Hockney, MIMD computing in the USA-1984, Parallel Computing, vol 2 (1985), pp. 119–136
U. Haas, Modelling of a program by an artificial benchmark program for vector computers, with discussion of the efficiency of the veetorization, Interner Berich Nr. 31/87 des Rechenzentrums der Universität Karlsruhe, 1987. Free copies of this internal report can be obtained on request.
Author information
Authors and Affiliations
Editor information
Editors and Affiliations
Rights and permissions
Copyright information
© 1989 Springer-Verlag Berlin Heidelberg
About this paper
Cite this paper
Schönauer, W. (1989). Why I like Vector Computers. In: Meuer, H.W. (eds) Supercomputer ’89. Informatik-Fachberichte, vol 211. Springer, Berlin, Heidelberg. https://doi.org/10.1007/978-3-642-74844-8_10
Download citation
DOI: https://doi.org/10.1007/978-3-642-74844-8_10
Publisher Name: Springer, Berlin, Heidelberg
Print ISBN: 978-3-540-51310-0
Online ISBN: 978-3-642-74844-8
eBook Packages: Springer Book Archive