I/O requirements of scientific applications: an evolutionary view

The modest I/O configurations and file system limitations of many current high-performance systems preclude solution of problems with large I/O needs. I/O hardware and file system parallelism is the key to achieving high performance. We analyze the I/O behavior of several versions of two scientific applications on the Intel Paragon XP/S. The versions involve incremental application code enhancements across multiple releases of the operating system. Studying the evolution of I/O access patterns underscores the interplay between application access patterns and file system features. Our results show that both small and large request sizes are common, that at present, application developers must manually aggregate small requests to obtain high disk transfer rates, that concurrent file accesses are frequent, and that appropriate matching of the application access pattern and the file system access mode can significantly increase application I/O performance. Based on these results, we describe a set of file system design principles.

[1]  Andrew A. Chien,et al.  Performance Modeling of a Parallel I/O System: An Application Driven Approach , 1997, PPSC.

[2]  Garth A. Gibson,et al.  RAID: high-performance, reliable secondary storage , 1994, CSUR.

[3]  Michael J. Quinn,et al.  Data-parallel programming on a network of heterogeneous workstations , 1992, Proceedings of the First International Symposium on High-Performance Distributed Computing. (HPDC-1).

[4]  Giuseppe Serazzi,et al.  Performance Gains from Leaving Idle Processors in Multiprocessor Systems , 1995, ICPP.

[5]  D.A. Reed,et al.  Input/Output Characteristics of Scalable Parallel Applications , 1995, Proceedings of the IEEE/ACM SC95 Conference.

[6]  Rice UniversityCORPORATE,et al.  High performance Fortran language specification , 1993 .

[7]  D.A. Reed,et al.  Scalable performance analysis: the Pablo performance analysis environment , 1993, Proceedings of Scalable Parallel Libraries Conference.

[8]  Evgenia Smirni,et al.  Thread placement on the Intel Paragon: modeling and experimentation , 1995, MASCOTS '95. Proceedings of the Third International Workshop on Modeling, Analysis, and Simulation of Computer and Telecommunication Systems.

[9]  Evgenia Smirni,et al.  The KSR1: experimentation and modeling of poststore , 1993, SIGMETRICS '93.

[10]  Carla Schlatter Ellis,et al.  File-Access Characteristics of Parallel Scientific Workloads , 1996, IEEE Trans. Parallel Distributed Syst..

[11]  Evgenia Smirni,et al.  Measuring the Effects of Thread Placement on the Kendall Square KSR1 , 1993 .

[12]  Randy H. Katz,et al.  Input/output behavior of supercomputing applications , 1991, Proceedings of the 1991 ACM/IEEE Conference on Supercomputing (Supercomputing '91).

[13]  C. Winstead,et al.  Parallel computation of electron molecule collisions , 1995 .

[14]  G. C. Polyzos,et al.  A static analysis of I/O characteristics of scientific applications in a production workload , 1993, Supercomputing '93.

[15]  Ken Kennedy,et al.  Software support for irregular and loosely synchronous problems , 1992 .

[16]  Andrew A. Chien,et al.  Input/Output Characteristics of Scalable Parallel Applications , 1995, SC.

[17]  Dror G. Feitelson,et al.  Parallel access to files in the Vesta file system , 1993, Supercomputing '93. Proceedings.

[18]  N. S. Barnett,et al.  Private communication , 1969 .

[19]  Giuseppe Serazzi,et al.  Analysis of Non-Work-Conserving Processor Partitioning Policies , 1995, JSSPP.

[20]  Peter Steenkiste,et al.  Automatic generation of parallel programs with dynamic load balancing , 1994, Proceedings of 3rd IEEE International Symposium on High Performance Distributed Computing.

[21]  J. T. Poole.  Preliminary survey of i/o intensive applications , 1994 .

[22]  D. G. Feitelson,et al.  Parallel access to files in the Vesta file system , 1993, Supercomputing '93.

[23]  Sanjay Ranka,et al.  Partitioning unstructured computational graphs for nonunifor , 1995, IEEE Parallel & Distributed Technology: Systems & Applications.

[24]  Sanjay Ranka,et al.  Architecture-independent locality-improving transformations of computational graphs embedded in k-dimensions , 1995, ICS '95.

[25]  Johan De Keyser,et al.  Run-Time Load Balancing Support for a Parallel Multiblock Euler/Navier-Stokes Code with Adaptive Refinement on Distributed Memory Computers , 1994, Parallel Comput..

[26]  Evgenia Smirni,et al.  Performance visualization: 2-D, 3-D, and beyond , 1996, Proceedings of IEEE International Computer Performance and Dependability Symposium.

[27]  Evgenia Smirni,et al.  I/O, performance analysis, and performance data immersion , 1996, Proceedings of MASCOTS '96 - 4th International Workshop on Modeling, Analysis and Simulation of Computer and Telecommunication Systems.

[28]  Richard Wheeler,et al.  it/sfs: A Parallel File System for the CM-5 , 1993, USENIX Summer.

[29]  Evgenia Smirni,et al.  The next frontier: interactive and closed loop performance steering , 1996, 1996 Proceedings ICPP Workshop on Challenges for Parallel Processing.

[30]  Stephen E. Deering,et al.  Multicast routing in datagram internetworks and extended LANs , 1990, TOCS.

[31]  Harry Berryman,et al.  Execution time support for adaptive scientific algorithms on distributed memory machines , 1991, Concurr. Pract. Exp..

[32]  David Kotz,et al.  Dynamic file-access characteristics of a production parallel scientific workload , 1994, Proceedings of Supercomputing '94.

[33]  Andrew A. Chien,et al.  PPFS: a high performance portable parallel file system , 1995, ICS '95.

[34]  George C. Polyzos,et al.  Dynamic I/O characterization of I/O intensive scientific applications , 1994, Proceedings of Supercomputing '94.

[35]  Philip K. McKinley,et al.  Communication issues in parallel computing across ATM networks , 1994, IEEE Parallel & Distributed Technology: Systems & Applications.

[36]  Carla Schlatter Ellis,et al.  Characterizing parallel file-access patterns on a large-scale multiprocessor , 1995, IPPS.

[37]  Andrew A. Chien,et al.  Algorithmic influences on I/O access patterns and parallel file system performance , 1997, Proceedings 1997 International Conference on Parallel and Distributed Systems.

[38]  Ramon Puigjaner,et al.  Modeling Techniques and Tools for Computer Performance Evaluation , 1989, Springer US.

[39]  George Em Karniadakis,et al.  Unstructured spectral element methods for simulation of turbulent flows , 1995 .