An Experimental Evaluation of I/O Optimizations on Different Applications

ABSTRACT

Many large scale applications have significant I/O requirements as well as computational and memory requirements. Unfortunately, the limited number of I/O nodes provided in a typical configuration of the modern message-passing distributed-memory architectures such as Intel Paragon and IBM SP-2 limits the I/O performance of these applications severely. In this paper, we examine some software optimization techniques and evaluate their effects in five different I/O-intensive codes from both small and large application domains. Our goals in this study are twofold. First, we want to understand the behavior of large-scale data-intensive applications and the impact of I/O subsystems on their performance and vice versa. Second, and more importantly, we strive to determine the solutions for improving the applications' performance by a mix of software techniques. Our results reveal that different applications can benefit from different optimizations. For example, we found that some applications benefit from file layout optimizations whereas others take advantage of collective I/O. A combination of architectural and software solutions is normally needed to obtain good I/O performance. For example, we show that with a limited number of I/O resources, it is possible to obtain good performance by using appropriate software optimizations. We also show that beyond a certain level, imbalance in the architecture results in performance degradation even when using optimized software, thereby indicating the necessity of an increase in I/O resources.

REFERENCES

• 1: Meenakshi Arunachalam , Alok Choudhary, A prefetching prototype for the parallel file systems on the Paragon, Proceedings of the 1995 ACM SIGMETRICS joint international conference on Measurement and modeling of computer systems, p.321-322, May 15-19, 1995, Ottawa, Ontario, Canada
• 2: R. Bennett, K. Bryant, A. Sussman, R. Das, and J. Saltz, "Jovian: A Framework for Optimizing Parallel I/O," Proc. 1994 Scalable Parallel Libraries Conf. 1994.
• 3: Rajesh Bordawekar , Alok Choudhary , Ken Kennedy , Charles Koelbel , Michael Paleczny, A model and compilation strategy for out-of-core data parallel programs, Proceedings of the fifth ACM SIGPLAN symposium on Principles and practice of parallel programming, p.1-10, July 19-21, 1995, Santa Barbara, California, United States
• 4: Rajesh Bordawekar , Alok Choudhary , J. Ramanujam, Automatic optimization of communication in compiling out-of-core stencil codes, Proceedings of the 10th international conference on Supercomputing, p.366-373, May 25-28, 1996, Philadelphia, Pennsylvania, United States [doi>10.1145/237578.237638]
• 5: Peter Brezany , Thomas A. Mück , Erich Schikuta, Language, compiler and parallel database support for I/O intensive applications, Proceedings of the International Conference and Exhibition on High-Performance Computing and Networking, p.14-20, May 03-05, 1995
• 6: P.H. Carns, W.B. Ligon Ill, R.B. Ross, and R. Thakur, "PVFS: A Parallel File System for Linux Clusters," Preprint /INL/MCS-P804-0400, submitted to the 2000 Extreme Linux Workshop April 2000.
• 7: Yuqun Chen , James S. Plank , Kai Li, CLIP: a checkpointing tool for message-passing parallel programs, Proceedings of the 1997 ACM/IEEE conference on Supercomputing (CDROM), p.1-11, November 15-21, 1997, San Jose, CA [doi>10.1145/509593.509626]
• 8: A. Choudhary, R. Bordawekar, S. More, K. Sivaram, and R. Thakur, "The PASSION Runtime Library for the Intel Paragon," Proc. Intel Supercomputer User's Group Conf., June 1995.
• 9: P. Corbett, D. Feitelson, S. Fineherg, Y. Hsu, B. Nitzherg, J, Prost, M. Snir, B. Traversal, and P. Wong, "Overview of the MPI-10 Parallel I/O Interface," Pro. Third Workshop I/O in Parallel unit Distributed Systems, Apr. 1995.
• 10: Phyllis E. Crandall , Ruth A. Aydt , Andrew A. Chien , Daniel A. Reed, Input/output characteristics of scalable parallel applications, Proceedings of the 1995 ACM/IEEE conference on Supercomputing (CDROM), p.59-es, December 04-08, 1995, San Diego, California, United States [doi>10.1145/224170.224396]
• 11: P. F. Corbett , D. G. Feltelson , J.-P. Prost , G. S. Almasi , S. J. Baylor , A. S. Bolmarcich , Y. Hsu , J. Satran , M. Snir , R. Colao , B. D. Herr , J. Kavaky , T. R. Morgan , A. Ziotek, Parallel file systems for the IBM SP computers, IBM Systems Journal, v.34 n.2, p.222-248, 1995
• 12: J. Del Rosario, R. Bordawekar, and A. Choudhary, "Improved Parallel I/O via A Two-Phase Run-Time Access Strategy," Proc. 1993 IPPS Workshop Input/Output in Parallel Computer Sys tems, Apr. 1993.
• 13: Juan Miguel del Rosario , Alok N. Choudhary, High-performance I/O for massively parallel computers: problems and prospects, Computer, v.27 n.3, p.59-68, March 1994 [doi>10.1109/2.268887]
• 14: Samuel A. Fineberg, Implementing the NHT-1 application I/O benchmark, ACM SIGARCH Computer Architecture News, v.21 n.5, p.23-30, Dec. 1993 [doi>10.1145/165660.165665]
• 15: James V. Huber, Jr. , Andrew A. Chien , Christopher L. Elford , David S. Blumenthal , Daniel A. Reed, PPFS: a high performance portable parallel file system, Proceedings of the 9th international conference on Supercomputing, p.385-394, July 03-07, 1995, Barcelona, Spain [doi>10.1145/224538.224638]
• 16: Meenakshi Arunachalam Kandaswamy , Alok Choudhary, Design and evaluation of optimizations in i/o-intensive applications, 1998
• 17: Meenakshi A. Kandaswamy , Mahmut T. Kandemir , Alok N. Choudhary , David E. Bernholdt, Optimization and evaluation of Hartree-Fock application's I/O with PASSION, Proceedings of the 1997 ACM/IEEE conference on Supercomputing (CDROM), p.1-20, November 15-21, 1997, San Jose, CA [doi>10.1145/509593.509624]
• 18: Mahmut T. Kandemir, A Collective I/O Scheme Based on Compiler Analysis, Selected Papers from the 5th International Workshop on Languages, Compilers, and Run-Time Systems for Scalable Computers, p.1-15, May 25-27, 2000
• 19: M. Kandemir , A. Choudhary , J. Ramanujam , R. Bordawekar, Compilation techniques for out-of-core parallel computations, Parallel Computing, v.24 n.3-4, p.597-628, May, 1998 [doi>10.1016/S0167-8191(98)00027-1]
• 20: M. Kandemir , A. Choudhary , J. Ramanujam , M. Kandaswamy, A unified compiler algorithm for optimizing locality, parallelism and communication in out-of-core computations, Proceedings of the fifth workshop on I/O in parallel and distributed systems, p.79-92, November 17-17, 1997, San Jose, California, United States [doi>10.1145/266220.266228]
• 21: Mahmut T. Kandemir , J. Ramanujam , Alok N. Choudhary, Improving the Performance of Out-of-Core Computations, Proceedings of the international Conference on Parallel Processing, p.128-136, August 11-15, 1997
• 22: D. Kotz, Expanding the potential for disk-directed I/O, Proceedings of the 7th IEEE Symposium on Parallel and Distributeed Processing, p.490, October 25-28, 1995
• 23: David Kotz , Carla Schlatter Ellis, Practical prefetching techniques for multiprocessor file systems, Distributed and Parallel Databases, v.1 n.1, p.33-51, Jan. 1993 [doi>10.1007/BF01277519]
• 24: Todd C. Mowry , Angela K. Demke , Orran Krieger, Automatic compiler-inserted I/O prefetching for out-of-core applications, Proceedings of the second USENIX symposium on Operating systems design and implementation, p.3-17, October 29-November 01, 1996, Seattle, Washington, United States
• 25: "NWChem, A Computational Chemistry Package for Parallel Computers, Version 1. 1,"High Performance Computational Chemistry Group, Pacific Northwest Laboratory (PNL, 1995.
• 26: M, Paleczny, K. Kennedy, and C. Koelbel, "Compiler Support for Out-of-Core Arrays on Parallel Machines,," CRI'C Technical Report 94509-S. Rice Univ., Houston l ex,, Dec. 1994,
• 27: D. Reed, R. Aydt, R. Noe, P. Ruth, K. Shields, B. Schwartz, and L. Tavera, "Scalable Performance Analysis'. the Pablo Performance Analysis Environment," Proc. scalable Parallel Libraries Cool., pp. 104-113, 1993.
• 28: B. Rullman Paragon Parallel File System, External Product Specification, Intel Supercomputer Systems Division. 1996.
• 29: 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]
• 30: F. Smirni, C. Elford, A, Laevry, D. Reed, and A. Chien, "Algorithmic Influences on I/O Access Patterns and Parallel File System Performance," Technical Report, Table Group, Univ. of Illinois at Urbana-Champaign, 1996.
• 31: Rajeev Thakur , Alok Choudhary, An extended two-phase method for accessing sections of out-of-core arrays, Scientific Programming, v.5 n.4, p.301-317, Winter 1996
• 32: Rajeev Thakur , William Gropp , Ewing L. Lusk, An Experimental Evaluation of the Parallel I/O Systems of the IBM SP and Intel Paragon Using a Production Application, Proceedings of the Third International ACPC Conference with Special Emphasis on Parallel Databases and Parallel I/O: Parallel Computation, p.24-35, September 23-25, 1996
• 33: R. Thakur, W. Cropp, and E. Lusk, "A Case for Using MPI's Derived Data Types to Improve I/O Performance, Preprint, ANL/ MCS-P717-0598, Math. and Computer Science Division, Argonne Nat'l Laboratory, May 1998,
• 34: Sivan Toledo , Fred G. Gustavson, The design and implementation of SOLAR, a portable library for scalable out-of-core linear algebra computations, Proceedings of the fourth workshop on I/O in parallel and distributed systems: part of the federated computing research conference, p.28-40, May 27-27, 1996, Philadelphia, Pennsylvania, United States [doi>10.1145/236017.236029]

INDEX TERMS

Primary Classification:
  B. Hardware
  B.4 INPUT/OUTPUT AND DATA COMMUNICATIONS
      B.4.2 Input/Output Devices

Additional Classification:
  B. Hardware
  B.8 Performance and Reliability
      B.8.2 Performance Analysis and Design Aids

  D. Software
  D.4 OPERATING SYSTEMS
      D.4.4 Communications Management
          Subjects: Input/output

General Terms:
Experimentation, Performance

Keywords:
I/O optimizations, parallel architectures, I/O intensive applications, disk layout, collective I/O.