|
| This Article | ||
| ||
| Share | ||
| Bibliographic References | ||
| Add to: | ||
 Digg  Furl  Spurl  Blink  Simpy  Google  Del.icio.us  Y!MyWeb | ||
| Search | ||
| ||
| ASCII Text | x | ||
| N. Islam, R.H. Campbell, "Design Considerations for Shared Memory Multiprocessor Message Systems," IEEE Transactions on Parallel and Distributed Systems, vol. 3, no. 6, pp. 702-711, November, 1992. | |||
| BibTex | x | ||
| @article{ 10.1109/71.180625, author = {N. Islam and R.H. Campbell}, title = {Design Considerations for Shared Memory Multiprocessor Message Systems}, journal ={IEEE Transactions on Parallel and Distributed Systems}, volume = {3}, number = {6}, issn = {1045-9219}, year = {1992}, pages = {702-711}, doi = {http://doi.ieeecomputersociety.org/10.1109/71.180625}, publisher = {IEEE Computer Society}, address = {Los Alamitos, CA, USA}, } | |||
| RefWorks Procite/RefMan/Endnote | x | ||
| TY - JOUR JO - IEEE Transactions on Parallel and Distributed Systems TI - Design Considerations for Shared Memory Multiprocessor Message Systems IS - 6 SN - 1045-9219 SP702 EP711 EPD - 702-711 A1 - N. Islam, A1 - R.H. Campbell, PY - 1992 KW - Index Termscomparative performance; message passing system designs; shared memory EncoreMultimax multiprocessor; benchmarks; example parallel applications; shared memorymachine results; Intel iPSC/2; NX/2 operating system; design alternatives; bufferorganization; value semantics; synchronization; coordination strategy; kernel space;memory caching; message sizes; copying; message passing; parallel machines; parallelprogramming; program testing; shared memory systems VL - 3 JA - IEEE Transactions on Parallel and Distributed Systems ER - | |||
The comparative performance is studied of different message passing system designsexperimentally on a shared memory Encore Multimax multiprocessor. The systems aremeasured both by benchmarks and by running example parallel applications. To act as acontrol, the shared memory machine results are compared with the performance of thebenchmarks and applications on the Intel iPSC/2 running the NX/2 operating system. Thedesign alternatives considered are buffering, buffer organization, reference and valuesemantics, synchronization, coordination strategy and the location of the system in useror kernel space. The results include measurements of the effects of the designalternatives, memory caching, message sizes and copying.
[1] T. E. Anderson, "The performance of spinlocking alternatives for shared memory multiprocessors,"IEEE Trans. Parallel Distributed Syst., pp. 6-16, Jan. 1990.
[2] B. Bershad, T. Anderson, E. Lazowska, and H. Levy, "User level interprocess communication for shared memory multiprocessors,"ACM Trans. Comput. Syst., vol. 9, no. 2, pp. 175-198, May 1991.
[3] R. H. Campbell, V. Russo, and G. Johnston, "Choices: The design of a multiprocessor operating system," inProc. USENIX C++ Workshop, Santa Fe, NM, Nov. 1987, pp. 109-123.
[4] P. Close, "The iPSC/2 Node Architecture,"Proc. Third Conf. Hypercube Concurrent Computers and Applications, Jan. 1988, SIAM.
[5] Encore Computer Corp.,Multimax Technical Summary, Encore, Marlborough, MA, 1986.
[6] J.R. Goodman, M.K. Vernon, and P.J. Woest, "Efficient Synchronization Primitives for Large-Scale Cache-Coherent Multiprocessors,"Proc. Third Int'l Conf. Architectural Support for Programming Languages and Operating Systems, IEEE CS Press, Los Alamitos, Calif., Order No. 1936, 1989, pp. 64-73.
[7] A. Gottlieb, B. D. Lubachevsky, and L. Rudolph, "Basic techniques for the efficient coordination of very large numbers of cooperating sequential processors,"ACM Trans. Programming Languages Syst., vol. 5, no. 2, pp. 164-189, Apr. 1993.
[8] N. Islam and R. Campbell, "Design considerations for shared memory multiprocessor message systems," Tech. Rep. UIUCDCS-R-91-1764, Univ. of Illinois Urbana-Champaign, Dec. 1991.
[9] H. Jacobs, "A user-tunable multiple processor scheduler," in1986 Winter USENIX Conf. Proc., Jan. 1986, pp. 183-191.
[10] T. J. LeBlanc, "Shared memory versus message passing in a tightly coupled multiprocessor: A case study," inProc. 1986 Int. Conf. Parallel Processing, ACM and IEEE, 1986, pp. 463-466.
[11] C. Lin and L. Snyder, "A comparison of programming models for shared memory multiprocessors," inProc. 1990 Int. Conf. Parallel Processing, Aug. 1990, pp. 163-170.
[12] A. Malony, D. Reed, and Rudolph, "Integrating Performance Data Collection, Analysis, and Visualization," inParallel Computer Systems: Performance Instrumentation and Visualization, M. Simmons, R. Koskela, and I. Bucher, eds., ACM, New York, 1990.
[13] H. Miyata, T. Isonishi, and A. Iwase, "Fast Fourier transformation using cellular array processor," inProc. Parallel Processing Symp. JSPP 1989, Feb. 1989, pp. 297-304.
[14] J.C. Mogul and A. Borg, "The Effects of Context Switches on Cache Performance,"Proc. ASPLOS-IV, Apr. 1991, pp. 75-84.
[15] J. K. Ousterhout, "Scheduling techniques for concurrent systems," inProc. Third Int. Conf. Distributed Comput. Syst., July 1982, pp. 22-30.
[16] P. Pierce, "The NX/2 Operating System,"Proc. Third Conf. Hypercube Concurrent Computers and Applications, ACM Press, 1988, pp. 384-390.
[17] V. F. Russo, P. W. Madany, and R. H. Campbell, "C++ and operating systems performance: A case study," inProc. USENIX C++ Conf., San Francisco, CA, Apr. 1990, pp. 103-114.
[18] C. B. Stunkel, "Linear optimization via message-based parallel processing," inProc. 1988 Int. Conf. Parallel Processing, Aug. 1988, pp. 264-271.
[19] J. Zahorjan, E. Lazowska, and D. Eager, "Spinning versus blocking in parallel systems with uncertainty," inProc. Int. Symp. Performance of Distributed and Parallel Syst., Dec. 1988.
[20] J. Zahorjan, E. Lazowska, and D. Eager, "The effect of scheduling discipline on spin overhead on shared memory parallel systems,"IEEE Trans. Parallel Distributed Syst., pp. 180-198, Apr. 1991.