Relationship between average and real memory behavior
Rent the article at a discountRent now
* Final gross prices may vary according to local VAT.Get Access
The CRAY Y-MP has a nonintrusive hardware performance monitor that accurately accumulates certain data about program performance. This paper examines the relationship between the averages obtained from the hardware performance monitor and actual memory behavior of the Perfect Club Benchmarks run on a single processor of an eight-processor CRAY Y-MP. I/O and instruction buffer fetches are not considered. The vectorized programs show regular behavior characterized by dominant vector lengths and interburst times. The distribution of vector lengths is not well-predicted by hardware performance monitor averages. Scalar programs also exhibit some clumping of memory references but have less temporal regularity than the vectorized programs. While overall port utilization is surprisingly low, there is considerable cyclic variation, and all of the ports tend to experience their maximal loading at the same time. A simple probabilistic model is developed to allow estimation of port utilitzation from hardware performance monitor data. The results can be used as a guide for generating more realistic synthetic memory workloads and port utilization estimates for shared-memory machines.
- Bailey, D.H. 1987. Vector computer memory bank contention.IEEE Trans. Comps., C-36: 293–298.
- Berry, M., Cybenko, G., and Larson, J. 1991. Scientific benchmark characterization.Parallel Comp., 17: 1173–1194.
- Bradley, D.K., and Larson, J.L. 1993. A parallelism-based analytic approach to performance evaluation using applications programs.IEEE Proc. 81: 1126–1135.
- Bradley, D., Cybenko, G., Gao, H., Larson, J., Ahmad, F., Golab, J., and Straka, M. 1991. Supercomputer workload decomposition and analysis. InProc., ACM Internat. Conf. on Supercomputing, pp. 458–467.
- Briggs, F.A., and Davidson, E.S. 1977. Organization of semiconductor memories for parallel-pipelined processors.IEEE Trans. Comps., C-26; 162–169.
- Cybenko, G. 1991. Supercomputer performance trends and the Perfect Benchmarks.Supercomputing Review (Apr.): 53–60.
- Gao, H., and Larson, J.L. 1993. A year's profile of academic supercomputer users using the Cray hardware performance monitor. Center for Supercomputing Res. and Dev., CSRD Rept. 1300, Urbana, Ill.
- Larson, J.L. 1991. Collecting and interpreting hpm performance data on the Cray Y-MP.NCSA Datalink (Nov.–Dec.).
- Malony, A.D., Larson, J.L., and Reed, D.A. 1990. Tracing application program execution on the Cray X-MP and Cray 2. InProc., Supercomputing '90, pp. 60–73.
- Robbins, K.A., and Robbins, S. 1991. Bus conflicts for logical memory banks on a Cray Y-MP type processor system. InProc., Internat. Conf. on Parallel Processing, pp. 21–24.
- Robbins, K.A., and Robbins, S. 1992. Experimental assessment of the Perfect Club Benchmarks on a Cray Y-MP. Univ. of Tex. at San Antonio, Tech. Rept. UTSA-CS-92-102.
- Robbins, K.A., and Robbins, S. 1994. Buffered banks in multiprocessor systems.IEEE Trans. Comps. (to appear).
- Vajapeyam, S., Sohi, G.S., and Hsu, W.-C. 1991. An empirical study of the Cray Y-MP processor using the Perfect Club Benchmarks. InProc., 18th Internat. Symp. on Comp. Architecture, pp. 170–179.
- Williams, E., Myers, C.T., and Koskela, R. 1990. The characterization of two scientific workloads using the Cray X-MP performance monitor. InProc., Supercomputing '90, pp. 142–152.
- Relationship between average and real memory behavior
The Journal of Supercomputing
Volume 8, Issue 3 , pp 209-232
- Cover Date
- Print ISSN
- Online ISSN
- Kluwer Academic Publishers
- Additional Links
- Hardware performance monitor
- Perfect Club
- CRAY Y-MP
- memory utilitzation
- shared memory
- Industry Sectors