Performance Attributes for Code and Workload Analysis On Cray X-Mp and Y-Mp Systems

Abstract

A Performance Analysis Tools (PAT) report implement ing hpm and perftrace software, installed under UNICOS on CRAY X-MP and Y-MP systems, was de signed at the Ohio Supercomputer Center. The report presents a concise summary of code performance characteristics using 28 performance attributes grouped into four categories: operations and instruc tions, scalar to vector load balance, memory utilization, and input/output utilization. Demonstrated applications of this performance metric include code performance analysis in a case study, workload analysis for a 24- code sample, and performance comparison of CRAY X-MP and Y-MP processors. The analysis of the case study shows that a precise performance assessment was possible using this metric without the necessity of source code analysis. Measurements with the 24-code sample established realistic performance ranges in the metric and were also used to compare performance of CRAY X-MP and Y-MP models. Although the clock pe riod on the Y-MP was reduced by 29% relative to the X-MP, increases of 52% were observed in instruction fetch rates and vector MFLOPS, with a corresponding rise of 39% in computational intensity.

References

Berry, M. , Chen, D. , Koss, P. , et al. 1989. The Perfect Club benchmarks: effective performance evaluation of supercomputers. Internat. J. Super-comput. Appl . 3(3):5-40.

Berry, M. , Cybenko, G. , and Larson, J. 1991. Scientific benchmark characterizations. Parallel Comput. 17:1173-1194.

Bradley, D. , Cybenko, G. , Gao, H. , Ahmad, F. , Golab, J. , Larson, J. , and Straka, M. 1991. Supercomputer workload decomposition and analysis . In Proceedings of the ACM international conference on supercomputing . New York: ACM Press, pp. 458-467.

Bradley, D.K. , and Larson, J.L. 1991. Fine-grained measurement of loop performance on the CRAY Y-MP. In Fifth SIAM conference on parallel processing for scientific applications, edited by D. C. Sorenson Philadelphia: SIAM.

Chen, S. , Hsiung, C. , Larson, J. , and Somdahl, E. 1988. CRAY X-MP: a multiprocessor supercomputer. In High speed and large-scale computing: a panoramic view, edited by

M. Ginsberg. Warrendale, Pennsylvania: Society of Automotive Engineers, pp. 57-70.

Cray Research, Inc. 1989. UNICOS performance utilities reference manual, SR-2040 version B (UNICOS 5.0).

Cray Research, Inc. 1990. CRAY Y-MP system programmer reference manual, CSM-2040-0A0.

Cray Research, Inc. Software Training. 1990 . CRAY Y-MP system architecture for applications programmers, TR-YSAAP.

10.

Cray Research, Inc. 1991a. UNICOS performance utilities reference manual, SR-2040 version 6.0.

11.

Cray Research, Inc. 1991b. CRAY Y-MP E series functional description manual, HR-04016-0A.

12.

Delic, G. 1991. Performance analysis of a 24 code sample on CRAY X/Y-MP systems at the Ohio Supercomputer Center. In Fifth SIAM conference on parallel processing for scientific applications, edited by D. C. Sorenson Philadelphia: SIAM.

13.

Elster, C. 1992. Microscopic calculations of nuclear scattering . Ohio SuPercomputer Center Visions Fall:11-15.

14.

Faanes, G.J. , and Schwarzmeier, J.L. 1989. Comparing the performance of CRAY Y-MF and CRAY X-MP computer systems. Cray Channels , Winter:26-30.

15.

Ginsberg, M. 1989. Computational environment influences on the performance of crashworthiness programs. In ASME AMD, vol 106. Crashworthiness and occupant protection in transportation systems, edited by T. B. Kahlil and A. I. King. pp. 11-21.

16.

Infante, E.F. 1986. An agenda for improved evaluation of supercomputer performance, a report prepared by the Committee on Supercomputer Performance and Development . Washington, D.C.: National Academy Press.

17.

Larson, J. 1986. CRAY X-MP hardware performance monitor. Cray Channels, Winter:18-19.

18.

Larson, J. , and Lutz, R. 1985. Perftrace user guide. Technical Report. Cray Research , Inc.

19.

Levesque, J.M. and Williamson, J.W. 1989. A guidebook to Fortran on supercomputers. New York: Academic Press.

20.

McMahon, F.H. 1986. The Livermore Fortran kernels: a computer test of the numerical performance range, UCRL-53745 Livermore, Calif ornia: Lawrence Livermore National Laboratory .

21.

National Science Foundation, Division of Advanced Scientific Computing. 1989. Report of the Workshop on Scientific Computing Performance Analysis, August 29-31, Boulder, Colorado.

22.

Patterson, D.A. , and Hennessy, J.L. 1990. Computer architecture: a quantitative approach . San Mateo, California: Morgan Kaufmann Publishers.

23.

Robbins, K.A. , and Robbins, S. 1990. The CRAY X-MP/Model 24: a case study in pipelined architecture and vector processing. New York: Springer-Verlag.

24.

Villumsen, J.V. 1989. A new hierarchical particle-mesh code for very large cosmological N-body simulations. Ap. J. Suppl. 71: 407-431.

25.

West, E. 1989. X-MP system wide performance analysis. Center for Large Scale Computation News (Ontario) 3(1):15-16.

26.

Williams, E. , Myers, C.T. , and Koskela, R. 1990. The characterization of two scientific workloads using the CRAY X-MP performance monitor. In Proceedings of Supercomputing '90. Los Alamitos, California: IEEE Computer Society Press , pp. 142-152.