This Article 
   
 Share 
   
 Bibliographic References 
   
 Add to: 
 
Digg
Furl
Spurl
Blink
Simpy
Google
Del.icio.us
Y!MyWeb
 
 Search 
   
MAD Kernels: An Experimental Testbed to Study Multiprocessor Memory System Behavior
February 1996 (vol. 7 no. 2)
pp. 207-217

Abstract—On large-scale multiprocessors, access to common memory is one of the key performance limiting factors. The shared-memory performance depends not only on the characteristics of the memory hierarchy itself, but also upon the characteristics of the memory address streams and the interaction between the two. We present a technique for multiprocessor workload construction and a family of artificial kernels, called MAD-kernels, to systematically investigate the behavior of the memory hierarchy. The measured performance is independent of any particular application or algorithm. The proposed methodology is demonstrated on two commercial shared-memory systems.

[1] K. Gallivan, D. Gannon, W. Jalby, A. Malony, and H. Wijshoff, "Behavioral characterization of multiprocessor memory systems: A case study," ACM Sigmetrics Performance Evaluation Review, vol. 17, pp. 79 - 88, May 1989.
[2] R.H. Saavedra, R.S. Gaines, and M.J. Carlton, "Micro Benchmark Analysis of the KSR1," Proc. Supercomputing 93, IEEE CS Press, 1993, pp. 202-213.
[3] A. Gupta and W.-D. Weber, "Cache Invalidation Patterns in Shared-Memory Multiprocessors," IEEE Trans. Computers, vol. 41, no. 7, pp. 794-810, July 1992.
[4] S. Baylor and B. Rathi, "A study of the memory reference behavior of engineering/scientific applications in parallel processors," Proc. 1989 Int'l Conf. on Parallel Processing, vol. 1, pp. 78 - 82, 1989.
[5] A. Gupta et al., "Comparative Evaluation of Latency Reducing and Tolerating Techniques," Proc. 18th Annual Int'l Symp. Computer Architecture, IEEE CS Press, Los Alamitos, Calif., June 1991, pp. 254-263.
[6] J. Torellas, M. Lam, and J. Hennessy, "Measurement, analysis, and improvement of the cache behavior of shared data in cache-coherent multiprocessors," Tech. Rep. CSL-TR-90-412, Stanford Univ., Feb. 1990.
[7] D.F. Vrsalovic, D.P. Siewiorek, Z.Z. Segall, and E.F. Gehringer, "Performance prediction and calibration for a class of multiprocessors," IEEE Trans. Computers, vol. 37, no. 11, pp. 1,353-1,364, Nov. 1988.
[8] R. Saavedra-Barrera, "Machine characterization and benchmark performance prediction," Tech. Rep. UCB/CSD 88/437, Univ. of California, Berkeley, June 1989.
[9] A. Nanda and L. M. Ni, "SAD kernels: A software tool to evaluate synchronization behavior of multiprocessors," Proc. 1992 Computer Science and Applications Conf., Sept. 1992.
[10] A. Nanda, "A framework for multiprocessor performance characterization and calibration,". PhD thesis, Michigan State Univ., East Lansing, Dept. of Computer Science, Oct. 1992.
[11] R. Hockney, "Performance characterization of the HEP," MIMD Computation: HEP Supercomputer and its Applications, J. Kowalik, ed. Cambridge, Mass: MIT Press, 1985.
[12] R. Hockney, "(r∞, n1/2, s1/2) measurements on the 2-CPU CRAY X-MP," Parallel Computing vol. 2, pp. 1-14, North-Holland, 1985.
[13] Sequent Computer Systems Inc., Symmetry Technical Summary, 1987.
[14] BBN Advanced Computers Inc., Cambridge, Mass., TC2000 Technical Product Summary, Nov. 1989.
[15] G. Pfister and V. Norton, "Hot-spot contention and combining in multistage interconnection networks," IEEE Trans. Computers, vol. 34, no. 10, pp. 943-948, Oct. 1985.
[16] T. H. Dunigan, "Kendall Square multiprocessor: Early experiences and performance," Tech. Rep. ORNL/TM-12065, Oak Ridge Nat'l Laboratory, Oak Ridge, Mar. 1992.

Index Terms:
Performance evaluation, shared-memory multiprocessors, memory hierarchy, interconnection networks, resource contention, synchronization overhead, memory access patterns.
Citation:
Arun K. Nanda, Lionel M. Ni, "MAD Kernels: An Experimental Testbed to Study Multiprocessor Memory System Behavior," IEEE Transactions on Parallel and Distributed Systems, vol. 7, no. 2, pp. 207-217, Feb. 1996, doi:10.1109/71.485509
Usage of this product signifies your acceptance of the Terms of Use.