This article is concerned with research on parallel pro cessing for chemically reactive supersonic flows. Effi ciency of autotasking and microtasking is examined for the problems of linear algebra and two-dimensional Euler equations with the finite element method. For microtasking, satisfactory improvements in speedup (2.63 for 4 CPUs and 1.67 for 2 CPUs) are achieved, with the load balance and the amount of parallelism being the dominant factors for the speedup. Autotask ing has advantages in providing automatic paralleliza tion, but it has been shown that in the solution of finite element Euler equations based on conventional algo rithms, parallelization by means of user's intervention is required for significant speedups (3.35 for 4 CPUs and 1.79 for 2 CPUs). A much better performance may be achieved with a combination of autotasking and microtasking. These results are verified from various example problems of linear algebra and supersonic diffuser flow field calculations with hydrogen-oxygen chemical reactions.
Get full access to this article
View all access options for this article.
References
1.
Chung, T.J., Kim, Y.M., and Sohn, J.L.1987. Finite element analysis in combustion phenomena . Internat. J. Numer. Meth. Fluids7:989-1012.
2.
Dongarra, J.J., Gustavson, F.C., and Karp, A.1984. Implementing linear algebra algorithm for dense matrices on a vector pipeline machine. SIAM Rev.26:91-112.
3.
Dongarra, J.J., and Hewitt, T.1986. Implementing dense linear algebra algorithm using multitasking on the CRAY X-MP. SIAM J. Sci. Statist. Comput.7(1):347-350.
4.
Dongarra, J.J., and Hinds, A.R.1979. Unrolling loops in FORTRAN. Software, practice and experience , vol. 9. New York: Wiley, pp. 219-226.
5.
Drummond, J.P., Hussaini, M.Y., and Zang, T.A.1985. Spectral method for modeling supersonic chemically reacting flow fields. AIAA paperAIAA—85-0302.
6.
Faber, V., and Lubeck, O.M.1987. Comment on the paper: parallel efficiency can be greater than unity. Parallel Computing4:209-210.
7.
Fong, K.W., and Jordan, T.L., 1977. Some linear algebraic algorithms and their performance on CRAY-1. Los Alamos National Laboratory Report LA-6774.
Hossfeld, F., Knecht, R., and Nagel, W.E.1989. Multitasking: experience with application on a CRAY X-MP . Parallel Computing12:259-283.
10.
Janben, R.1987. A note on super linear speed-up. Parallel Computing4:211-213.
11.
Jespersen, D.C., and Levit, C.1989. A computational fluid dynamics algorithm on a massively parallel computer. Internat. J. Supercomput. Appl.3(4):9-27.
12.
Johnson, G.M.1987. Parallel processing in fluid dynamics, application of parallel processing in fluid dynamics. FED, vol. 47. New York: ASME, pp. 1-8.
13.
Johnson, G.M.1989. Parallel processing for fluid dynamics application . In Finite element analysis in fluids. Huntsville : University of Alabama, pp. 1474-1478.
14.
Kratz, M.1984. Vectorized finite element stiffness generation: tuning the Noor-Lambiotte algorithm. Parallel Computing2:121-132.
15.
Levit, C., and Jespersen, D.1988. Explicit and implicit solution of the Navier-Stokes equations on a massively parallel computer. Presented at the Symposium on Advances and Trends in Structural Mechanics and Fluid Dynamics, Washington, DC.
16.
Misegades, K., and Krause, L.1987. Microtasking of fluid mechanics codes on the CRAY X-MP, application of parallel processing in fluid mechanics. FED, vol. 47. New York: ASME , pp. 19-25.
17.
Myczkowski, J., Bromley, M., and McCowan, D.1991. Extremely fast finite difference techniques for the Connection Machine. AIAA paperAIAA—91-0436.
18.
Nagel, W.E.1990a. Using multiple CPUs for problem solving: experiences in multitasking on the CRAY X-MP/48. Parallel Computing8:223-230.
19.
Nagel, W.E.1990b. Exploiting autotasking on a CRAY Y-MP: an improved software interface to multitasking. Parallel Computing8:225-233.
20.
Noor, A.K., and Lambiotte, J.J.1979. Finite element dynamic analysis on CDC STAR-100 Computer . Computers & Structures10:151-161.
21.
Oed, W., and Lange, O.1985. On the effective bandwidth of interleaved memories in vector processor systems. IEEE Trans. Comput.C-34(10): 949-957.
22.
Ostega, J.M.1988. The ijk forms of factorization methods: I. Vector computers . Parallel Computing7:135-147.
23.
Patel, N.R., Sturek, W.B., and Hiromoto, R.E.1987. A parallel compressible flow algorithm for multiprocessors, application for parallel processing in fluid dynamics. FED , vol. 47, New York : ASME, pp. 49-64.
24.
Roger, R.C., and Chinitz, W.1983. Using a global hydrogen-air combustion model in turbulent reacting flow calculations. AIAA J.21(4):586-592.
25.
Saati, A., Biringen, S., and Farhat, C.1990. Solving Navier-Stokes equation on a massively parallel processor: beyond the 1GFLOP performance. Internat. J. Supercomput. Appl.4(1):72-80.
26.
Shapiro, R.A.1991. Implementation of an Euler Navier-Stokes finite element algorithm on the Connection Machine. AIAA paper, AIAA—91-0438.
27.
Silvester, D.J.1988. Optimising finite element matrix calculations using the general technique of element vectorization. Parallel Computing8:157-164.
28.
Swisshelm, J.M.1989. A parallel finite difference method for computational aerodynamics. In Finite element analysis in fluids. Huntsville: University of Alabama, pp. 1486-1491.
29.
Yoon, C.S., and Chung, T.J.1990. Shock waves/turbulence interactions in reacting flows. AIAA paperAIAA—90-0359.
