Douglas Eadline

Doug...@Eadline.org

An abridged resume follows. If you are interested in a full resume, please contact me directly.

Who am I?

Experience

Why I like clusters

Prior to the advent of "Beowulf Clusters", their was no standard way to implement parallel software. Of course there was MPI, but other aspects of the parallel architecture were often very system specific. Indeed, there were even "cluster" parallel computers based on UNIX workstations.

When the Beowulf "concept" emerged, some things changed. First, hardware was based on commodity products and thus provided a low cost solution that was available from multiple sources. Second, the OS and software tools were open source and adaptable to many types of hardware. Suddenly, a very propriety and closed technology became open, flexible, and low cost. The number of parallel computers (Beowulf clusters) quickly grew and now provides a robust flexible platform on which to implement software.

We are now in an exciting time. Clusters have made parallel computing a reality for many people. Now the only thing that prevents us from creating parallel computing "killer applications" is our imagination. In addition, the advent of using graphics processors as parallel desktop supercomputers is very interesting to me as well.

Educational Background

Juniata College, Huntingdon, Pennsylvania. Received B.S. in Chemistry May, 1979.

Lehigh University, Bethlehem, Pennsylvania. Received Ph.D. in Analytical Chemistry September 1984.

Publications/Projects

• Linux Magazine (2005) monthly column "Cluster Rant"

• ClusterWorld Articles

  December 2003: Probing Gigabit Ethernet
  January 2004: An Interview with Steve Wallach,Assume Nothing, Test Everything
  February 2004: Hedging on Clusters: Clusters in Finance, The NAS Suite,
  May 2004: The ClusterWorld Benchmark Project June 2004: ClusterWorld Conference and Expo 2004: Part 1
  July 2004 ClusterWorld Conference and Expo 2004: Part 2
  August 2004 The ClusterWorld Benchmark Project FAQ
  September 2004;MySQL Cluster: Interview and Background,Winning with HPC: The HPC Users Conference
  November 2004: Benchmarking: The State of the CWBP
  January 2005: Ten Question for the EGA, SC2004 Wrap-up: HPC from Pittsburgh, The Value Cluster: Cheaper, Better, Faster
  March 2005: The Value Cluster Part 2: Software Installation
  May 2005: The Value Cluster Part Three: Packages and Testing
  

• Linux magazine, November 2004: Why Linux on Clusters?

• The Cluster Agenda

• The Cluster Monkey Benchmark Project

• Beowulf HOWTO,(www.tldp.org/HOWTO/Beowulf-HOWTO.html) Jacek Radajewski and Douglas Eadline, 1998 (now in archive)

• Cluster Quick Start, Douglas Eadline,1998

• BERT 77 Fortran Conversion, Sequential to Parallel FORTRAN conversion.

• Beowulf Performance Suite (BPS), A collection of tests to measure cluster performance.

• Basic Cluster Scripts A collection of scripts to help build and maintain a cluster.

Older Publications

• Maximizing "Return on Investment": Prolog Extensions for Effective Decision Support Applications, Anatholy F. Dedkov and Douglas J. Eadline, Practical Applications of Prolog, PAP 96.

• Performance Considerations for I/O-Dominant Applications on Parallel Computers, Anatholy F. Dedkov and Douglas J. Eadline, Proceedings of the International Conference on Parallel and Distributed Processing Techniques and Applications (PDPTA '95), Athens GA, 1995, H. R. Arabnia, ed.

• "The ?-a(X),b(Y),c(X,Y) Challenge",Anatholy F. Dedkov and Douglas J. Eadline, Internal Technical Paper, Paralogic, Inc.

• Implementing Prolog on Distributed Systems: n-parallel PROLOG, Douglas Eadline, ILPS '94, Post Conference Workshop on Design and Implementation of Parallel Logic Programming Systems, November 18, 1994, Ithaca NY, Enrico Pontelli, Gopal Gupta, Laboratory for Logic, DB, and Advanced Programming, New Mexico State University, Las Cruces, NM 88003.

• Design and Implementation of a Prolog-to-C Compiler, A.F. Dedkov and D.J. Eadline, Post Conference Workshop on ILPS '94, Implementation Techniques for Logic Programming Languages, November 17, 1994, Ithaca, NY, Koen De Bosschere, Bart Demoen, and Paul Tarau, ELIS, Universiteit Gent, St.-Pietersnieuwstraat 41, B-9000 Gent, Belgium.

• "How will you write software for a 1000 node parallel computer?", Parallelogram, #52, March 1993.

• "Mapping Search Graphs onto Arbitrary Transputer Networks", Douglas Eadline, Transputer Research and Applications 3, Alan S. Wagner ed., IOS Press, Washington, 1990.

• "Making PROLOG Parallel", Douglas J. Eadline, AI Expert, Vol. 4, No. 7, July 1989.

• "High Frequency Time Domain Dielectric Spectrometer for the Determination of Water in Polymer Coatings on Metal Substrates", Douglas J. Eadline and Henry Leidheiser, Jr., Rev. Sci. Instrum. (1985) 56, 1432.

• "Experimental Errors and Artifacts for a Time Domain Lumped Capacitor Dielectric Spectrometer", Douglas J. Eadline and Henry Leidheiser, Jr., Rev. Sci. Instrum. (1986) 57, 898.