Abstract
A three-dimensional global ocean circulation model has been modified to run on the BBN TC2000 multiple instruction stream/multiple data stream (MIMD) parallel computer. Two shared-memory parallel programming models have been used to implement the global ocean model on the TC2000: the TCF (TC2000 Fortran) fork-join model and the PFP (Parallel Fortran Preprocessor) split-join model. The method chosen for the parallelization of this global ocean model on a shared-memory MIMD machine is discussed. The performance of each version of the code has been measured by varying the processor count for a fixed-resolution test case. The statically scheduled PFP version of the code achieves a higher parallel computing efficiency than does the dynamically scheduled TCF version of the code. The observed differences in the performance of the TCF and PFP versions of the code are discussed. The parallel computing performance of the shared-memory implementation of the global ocean model is limited by several factors, most notably load imbalance and network contention. The experience gained while porting this large, “real world” application onto a shared-memory multiprocessor is also presented to provide insight to the reader who may be contemplating such an undertaking.
Similar content being viewed by others
References
Beeler, M. 1990.Inside the TC2000 Computer, First Release. BBN Advanced Computers, Inc., Cambridge, Mass., pp. 59–62.
Chervin, R.M. 1990. Private commun.
Crowther, W., Goodhue, J., Starr, E., Thomas, R., Milliken, W., and Blackadar, T. 1985. Performance measurements on a 128-node Butterfly parallel processor. InProc., 1985 Internat. Conf. on Parallel Processing (St. Charles, Ill., Aug. 20–23), IEEE Comp. Soc. Press, pp. 531–539.
Cvetanovic, Z. 1987. The effects of problem partitioning, allocation, and granularity on the performance of multipleprocessor systems.IEEE Trans. Comp., C-36, 4 (Apr.): 421–432.
Iqbal, M.A., Saltz, J.H., and Bokhari, S.H. 1986. Performance tradeoffs in static and dynamic load balancing strategies. InProc., 1986 Internat. Conf. on Parallel Processing (St. Charles, Ill., Aug. 19–22), IEEE Comp. Soc. Press, pp. 1040–1046.
Lapidus, L., and Pinder, G.F. 1982.Numerical Solution of Partial Differential Equations in Science and Engineering. Wesley-Interscience Pub., New York, pp. 385–421.
LeBlanc, T.J. 1988. Problem decomposition and communication tradeoffs in a shared-memory multiprocessor. InNumerical Algorithms for Modern Parallel Computer Achitectures, IMA Vols. in Mathematics and Its Applications, vol. 16, Springer-Verlag, pp. 145–162.
LeBlanc, T.J., Scott, M.L., and Brown, C.M. 1988. Large-scale parallel programming: Experience with the BBN Butterfly parallel processor.ACM/SIGPLAN Notices, 23, 9 (Sept.): 161–172.
Moy, B. 1989.Uniform System Programming in Fortran, Revision Beta 2.0 BBN Advanced Computers, Inc., Cambridge, Mass.
NERSC Documentation Group. 1990.Autotasking Document. Nat. Energy Reseach Supercomputer Center, Livermore, Calif.
Quigley, A. 1990.TC2000 FORTRAN Reference, Revision 2.0 BBN Advanced Compuers, Inc., Cambridge, Mass., pp. 331–384.
Rettberg, R., and Thomas, R. 1986. Contention is no obstacle to shared-memory multiprocessing.CACM, 29, 12 (Dec): 1202–1212.
Semtner, A.J. 1986. Finite-difference formulation of a world ocean model. InAdvaced Physical Oceanographic Numerical Modeling, Reidel Pub. Co., pp. 187–202.
Semtner, A.J., and Chervin, R.M. 1988. A simulation of the global ocean circulation with resolved eddies.J. Geophys. Res., 93, C12 (Dec.): 15,502–15,522.
Warren, K., Gorda, B., and Brooks, E.D. III. 1991. Programming in PFP. Rep. UCRL-MA-107028, Lawrence Livermore National Laboratory, Livermore, Calif.
Whitman S.R. 1991. Utilizing scalable shared memory multiprocessors for computer graphics rendering. Ph.D. diss., The Ohio State Univ., Columbus, Oh.
Author information
Authors and Affiliations
Rights and permissions
About this article
Cite this article
Procassini, R.J., Whitman, S.R. & Dannevik, W.P. Porting a global ocean model onto a shared-memory multiprocessor: Observations and guidelines. J Supercomput 7, 287–321 (1993). https://doi.org/10.1007/BF01206241
Received:
Accepted:
Issue Date:
DOI: https://doi.org/10.1007/BF01206241