Performance modeling for the panda array I/O library

ABSTRACT

We present an analytical performance model for Panda, a library for synchronized i/o of large multidimensional arrays on parallel and sequential platforms, and show how the Panda developers use this model to evaluate Panda's parallel i/o performance and guide future Panda development. The model validation shows that system developers can simplify performance analysis, identify potential performance bottlenecks, and study the design trade-offs for Panda on massively parallel platforms more easily than by conducting empirical experiments. More importantly, we show that the outputs of the performance model can be used to help make optimal plans for handling application i/o requests, the first step toward our long-term goal of automatically optimizing i/o request handling in Panda.

REFERENCES

• 1: Peter M. Chen , David A. Patterson, A new approach to I/O performance evaluation: self-scaling I/O benchmarks, predicted I/O performance, Proceedings of the 1993 ACM SIGMETRICS conference on Measurement and modeling of computer systems, p.1-12, May 10-14, 1993, Santa Clara, California, United States
• 2: Y. Chen , M. Winslett , K. E. Seamons , S. Kuo , Y. Cho , M. Subramaniam, Scalable message passing in Panda, Proceedings of the fourth workshop on I/O in parallel and distributed systems: part of the federated computing research conference, p.109-121, May 27-27, 1996, Philadelphia, Pennsylvania, United States [doi>10.1145/236017.236042]
• 3: Gregory R. Ganger , Yale N. Patt, The process-flow model: examining I/O performance from the system's point of view, Proceedings of the 1993 ACM SIGMETRICS conference on Measurement and modeling of computer systems, p.86-97, May 10-14, 1993, Santa Clara, California, United States
• 4: M. J. Clement , M. J. Quinn, Analytical performance prediction on multicomputers, Proceedings of the 1993 ACM/IEEE conference on Supercomputing, p.886-894, December 1993, Portland, Oregon, United States [doi>10.1145/169627.169856]
• 5: {Franke94} H. Franke, P. Hochschild, P. Pattnaik and M. Snir, MPI-F: An efficient implementation of MPI on IBM-SP1, International Conference on Parallel Processing, 1994.
• 6: {Feitelson95} D. Feitelson, P. Corbett, and J-P. Prost. Performance of the Vesta file system, IPPS, 1995.
• 7: {Kotz94} D. Kotz, Disk-directed i/o for MIMD multiprocessors, OSDI, 1994.
• 8: Daniel A. Menascé , Luiz André Barroso, A methodology for performance evaluation of parallel applications on multiprocessors, Journal of Parallel and Distributed Computing, v.14 n.1, p.1-14, Jan. 1992 [doi>10.1016/0743-7315(92)90093-3]
• 9: {MPI94} MPI: A Message-Passing Interface Standard, Message Passing Interface Forum, 1994.
• 10: {Reed93} D. Reed, R. Aydt, R. Noe, P. Roth, K. Shields, B. Schwartz and L. Tavera, Scalable performance analysis: The Pablo performance analysis environment, Scalable Parallel Libraries Conf., 1993.
• 11: K. E. Seamons , Y. Chen , P. Jones , J. Jozwiak , M. Winslett, Server-directed collective I/O in Panda, Proceedings of the 1995 ACM/IEEE conference on Supercomputing (CDROM), p.57-es, December 04-08, 1995, San Diego, California, United States [doi>10.1145/224170.224371]
• 12: R. H. Saavedra-Barrera , A. J. Smith , E. Miya, Machine Characterization Based on an Abstract High-Level Language Machine, IEEE Transactions on Computers, v.38 n.12, p.1659-1679, December 1989 [doi>10.1109/12.40845]
• 13: Stephen J. Von Worley , Alan J. Smith, Microbenchmarking and Performance Prediction for Parallel, University of California at Berkeley, Berkeley, CA, 1995

INDEX TERMS

Primary Classification:
  K. Computing Milieux
  K.6 MANAGEMENT OF COMPUTING AND INFORMATION SYSTEMS

Additional Classification:
  C. Computer Systems Organization
  C.5 COMPUTER SYSTEM IMPLEMENTATION

  D. Software
  D.4 OPERATING SYSTEMS

  F. Theory of Computation
  F.1 COMPUTATION BY ABSTRACT DEVICES

  I. Computing Methodologies
  I.6 SIMULATION AND MODELING

General Terms:
Design, Management, Performance, Theory