Server-directed collective I/O in Panda

ABSTRACT

We present the architecture and implementation results for Panda 2.0, a library for input and output of multidimensional arrays on parallel and sequential platforms. Panda achieves remarkable performance levels on the IBM SP2, showing excellent scalability as data size increases and as the number of nodes increases, and provides throughputs close to the full capacity of the AIX file system on the SP2 we used. We argue that this good performance can be traced to Panda's use of server-directed i/o (a logical-level version of disk-directed i/o [Kotz94b]) to perform array i/o using sequential disk reads and writes, a very high level interface for collective i/o requests, and built-in facilities for arbitrary rearrangements of arrays during i/o. Other advantages of Panda's approach are ease of use, easy application portability, and a reliance on commodity system software.

REFERENCES

• Bennett94: R. Bennett, K. Bryant, A. Sussman, R. Das, and J. Saltz, Jovian: A framework for optimizing parallel I/O, Proceedings of the 1994 Scalable Parallel Libraries Conference, pages 10-20. IEEE Computer Society Press, October 1994.
• Bordawekar93: R. Bordawekar , J. M. del Rosario , A. Choudhary, Design and Evaluation of primitives for Parallel I/O, Proceedings of the 1993 ACM/IEEE conference on Supercomputing, p.452-461, December 1993, Portland, Oregon, United States [doi>10.1145/169627.169782]
• Brezany95: 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
• Corbett94b: P. F. Corbett and D. G. Feitelson, Vesta file system programmer's reference. Technical Report Research Report RC 19898 (88058), IBM T.J. Watson Research Center, Yorktown Heights, NY 10598, October 1994. Version 1.01.
• Corbett95: P. Corbett, D. Feitelson, Y. Hsu, J. Prost, M. Snir, S. Fineberg, B. Nitzberg, B. Traversat, and P. Wong. MPI-IO: A Parallel File I/O Interface for MPI, Technical Report NAS-95-002, NASA Ames Research Center, January 1995.
• del Rosario94: J. M. del Rosario, M. Harry, A. Choudhary, The Design of VIP-FS: A Virtual, Parallel File System for High Performance Parallel and Distributed Computing, Technical Report SCCS-628, NPAC, Syracuse, NY, May 1994.
• Galbreath93: D. Levine , N. Galbreath , W. Gropp, Applications-driven parallel I/O, Proceedings of the 1993 ACM/IEEE conference on Supercomputing, p.462-471, December 1993, Portland, Oregon, United States [doi>10.1145/169627.169784]
• Karpovich94: John F. Karpovich , Andrew S. Grimshaw , James C. French, Extensible file system (ELFS): an object-oriented approach to high performance file I/O, Proceedings of the ninth annual conference on Object-oriented programming systems, language, and applications, p.191-204, October 23-28, 1994, Portland, Oregon, United States
• Kotz93b: David Kotz, Throughput of Existing Multiprocessor File Systems (An Informal Study), Dartmouth College, Hanover, NH, 1993
• Kotz94a: David Kotz , Nils Nieuwejaar, Dynamic file-access characteristics of a production parallel scientific workload, Proceedings of the 1994 conference on Supercomputing, p.640-649, December 1994, Washington, D.C., United States
• Kotz94b: D. Kotz, Disk-Directed I/O for MIMD Multiprocessors, First Symposium on Operating Systems Design and Implementation, November 1994.
• Kotz95a: David Kotz, Expanding the Potential for Disk-Directed I/O, Dartmouth College, Hanover, NH, 1995
• Kotz95b: David Kotz, Disk-directed I/O for an Out-of-Core Computation, Dartmouth College, Hanover, NH, 1995
• Pasquale94: B. Pasquale, and G. Polyzos, Dynamic I/O Characterization of I/O intensive Scientific Applications, Technical Report No. CS94-364, University of California, San Diego, April 1994.
• Pierce93: P. Pierce, A Concurrent File System for a Highly Parallel Mass Storage Subsystem, Proceedings of the 4th Conference on Hypercube Computers and Applications, Monterey, March 1989. , pp. 155-160.
• Purakayastha94: Apratim Purakayastha , Schlatter C Ellis , David Kotz , Nils Nieuwejaar , Michael Best, Characterizing Parallel File-Access Patterns on a Large-Scale Multiprocessor, Duke University, Durham, NC, 1994
• Seamons94a: Kent E. Seamons , Marianne Winslett, Physical Schemas for Large Multidimensional Arrays in Scientific Computing Applications, Proceedings of the Seventh International Working Conference on Scientific and Statistical Database Management, p.218-227, September 28-30, 1994
• Seamons94b: Kent E. Seamons , Marianne Winslett, An efficient abstract interface for multidimensional array I/O, Proceedings of the 1994 conference on Supercomputing, p.650-659, December 1994, Washington, D.C., United States
• Seligman94: Erik Seligman , Adam Beguelin, High-Level Fault Tolerance in Distributed Programs, Carnegie Mellon University, Pittsburgh, PA, 1994

INDEX TERMS

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

Additional Classification:
  C. Computer Systems Organization
  C.1 PROCESSOR ARCHITECTURES
      C.1.2 Multiple Data Stream Architectures (Multiprocessors)
          Subjects: Parallel processors**
  C.5 COMPUTER SYSTEM IMPLEMENTATION
      C.5.1 Large and Medium ("Mainframe") Computers
          Subjects: Super (very large) computers

General Terms:
Design, Performance