PACT'98 Final Program

Tuesday, October 13
09:00-17:30Workshop on Profile and Feedback-Directed Compilation
Workshop on Reconfigurable Computing
18:00-21:00Wine and Cheese Reception
Wednesday, October 14
09:25-09:30Welcoming remarks - Ulrich Finger, ENST, France, General co-chair
09:30-10:30 Keynote Speaker: Gordon Bell, Senior Researcher, Microsoft
Title: What Architectures? When? || Presentation slides
10:30-10:50Break
10:50-11:40Session 1: Loop transformations
Chair: Christine Eisenbeis, INRIA, France
Scanning Polyhedra without Do-loops || Presentation slides, P. Boulet, Laboratoire d'Informatique Fondamentale de Lille, France; P. Feautrier, Université de Versailles - St Quentin, France
Integrating Loop and Data Transformations for Global Optimisation, M. O'Boyle, Univ. of Edinburgh, UK; P. Knijnenburg, Leiden Univ., The Netherlands
11:40-11:55Break
11:55-13:10 Session 2: Shared memory design techniques
Chair: Nigel Topham, Univ. Edinburgh, UK
Optical versus electronic bus for Address-transactions in future SMP architectures, W. Hlayhel, D. Litaize, Université Paul Sabatier Toulouse, France; L. Fesquet, J. Collet, CNRS, Toulouse, France
Origin 2000 Design Enhancements for Communication Intensive Applications, G. Abandah, Univ. of Jordan; E. Davidson, Univ. of Michigan
Sirocco: Cost-Effective Fine-Grain Distributed Shared Memory, I. Schoinas, Intel; B. Falsafi, Purdue Univ.; M. Hill, J. Larus, D. Wood, Univ. of Wisconsin
13:10-14:25 Lunch break
14:25-15:30 Session 3: Specialized multiprocessor systems
Chair: Hidehiko Tanaka, University of Tokyo
Design Study of Shared Memory in VLIW Video Signal Processors, Z. Wu, W. Wolf, Princeton Univ.
Exploiting Fine- and Coarse-grain Parallelism in Embedded Programs, I. Karkowski, H. Corporaal, Delft Univ. of Technology, The Netherlands
Command Vector Memory Systems: High Performance at Low Cost, J. Corbal, R. Espasa, M. Valero, UPC-Barcelona, Spain
15:30-15:40 break
15:40-16:25 Session 4: Parallel programming languages
Chair: Paraskevas Evripidou, Univ. Cyprus
Eliminating Bottlenecks on Parallel Systems using Adaptive Objects, M. Yasugi, Kyoto Univ., Japan; S. Eguchi, K. Taki, Kobe Univ., Japan
Athapascan-1: On-line building data flow graph in a parallel language || Presentation slides, F. Gallilée, J.L. Roch, G. Cavalheiro, M. Doreille, Lmc-Imag Apache, France
General Parallel Computation Can be Performed with a Cycle-Free Heap, J. Dennis, MIT
16:25-16:40 Break
16:40-18:00 Session 5: Short papers
Chair: Jack Dennis, MIT
Code Generation in the Polytope Model || Presentation slides, M. Griebl, C. Lengauer, S. Wetzel, Univ. of Passau, Germany
Handling Cross Interferences by Cyclic Cache Line Coloring || Presentation slides, D. Genius, Univ. of Karlsruhe, Germany
Capturing the Effects of Code Improving Transformations || Presentation slides, C. Jaramillo, R. Gupta, M. Soffa, Univ. of Pittsburgh
Fast, Accurate and Flexible Data Locality Analysis, J. Sanchez, A. Gonzalez, UPC - Barcelona, Spain
Superscalar Execution With Direct Data Forwarding, S. Önder, R. Gupta, Univ. of Pittsburgh
A Fast Interrupt Handling Scheme for VLIW Processors, E. Özer, S. Sathaye, K. Menezes, S. Banerjia, M. Jennings, T. Conte, North Carolina State Univ.
Dynamic Elimination of Pointer-Expressions, N. Weinberg, D. Nagle, Carnegie Mellon Univ.
Efficacy and Performance Impact of Value Prediction, B. Rychlik, J. Faistl, B. Krug, J. Shen, Carnegie Mellon Univ.
18:00-19:00Short paper discussion and reception
Thursday, October 15
09:00-10:00 Keynote Speaker: Carole Dulong, Principal Engineer, IA-64 Architect, Intel
Title: "Next Generation ILP" || Presentation slides
10:00-10:20 Break
10:20-11:10 Session 6: ILP scheduling techniques
Chair: Uwe Schwiegelshohn, Univ. Dortmund, Germany
A fast algorithm for scheduling time-constrainted instructions on processors with ILP, A. Leung, K. Palem, New York Univ.; A. Pnueli, The Weizmann Institute of Science, Israel
A New Framework for Integrated Global Local Scheduling, S. Mantripragada, S. Jain, J. Dehnert, Silicon Graphics
11:10-11:25Break
11:25-12:05 Session 7: JAVA and multithreading multiprocessors
Chair: Jean-Luc Gaudiot, USC
Exploiting Method-Level Parallelism in Single-Threaded Java Programs, M. Chen, K. Olukotun, Stanford Univ.
The StarT-Voyager Parallel System, B. Ang, D. Chiou, L. Rudolph, Arvind, MIT
12:05-12:20Break
12:20-13:10 Session 8: Register allocation
Chair: Guang Gao, Univ. Delaware
Optimistic Register Coalescing J. Park, S. Moon, Seoul National Univ., Korea
Efficient JavaVM Just-in-Time Compilation, A. Krall, Technische Universität Wien, Austria
13:10-14:25 Lunch break
14:25-15:10 Session 9: Branch and value prediction
Chair: Mateo Valero, UPC Barcelona, Spain
Improving Static Branch Prediction in a Compiler, B. Deitrich, Motorola; B. Cheng, W. Hwu, Univ. of Illinois
Static Methods in Hybrid Branch Prediction, D. Grunwald, D. Lindsay, B. Zorn, Univ. of Colorado
Split Last-Address Predictor || Presentation slides, E. Morancho, J. Llaberia, A. Olive, Universidad Politecnica de Catalunya, Spain
15:10-15:20 Break
15:20-16:30 Session 10: Message passing systems
Chair: Claude Girault, Univ. Paris VI, France
Madeleine: An Efficient and Portable Communication Interface for RPC-based Multithreaded Environments || Presentation slides, L. Bougé, ENS Lyon, France; J. Méhaut, Université de Lille I, France; R. Namyst, ENS Lyon, France;
Breaking the Barriers: Two Models for MPI Programming J. Roda, C. Rodriguez, D. Morales, F. Almeida, P. Pulido, D. Dorta, Universidad de la Laguna, Spain
Adaptive Receiver Notification for Non-Dedicated Workstation Clusters, M. Schmitt, M. Ibel, A. Acharya, K. Schauser, Univ. of California at Santa Barbara
Performance of Message-Passing Systems Using a Zero-Copy Communication Protocol P. Melas, E. Zaluska, Univ. of Southampton, UK
16:30-16:45 Break
16:45-18:15 Panel: What should be the role of government intervention in advancing high performance computing? Views from around the world.

Panelists:

  • Dr. Erik Altman (chair), IBM Thomas J. Watson Research Center
  • Prof. Peter Behr, Director, GMD-FIRST, German National Research Center for Information Technology, Institute for Computer Architecture and Software Technology
  • Prof. Jack B. Dennis, M.I.T. Laboratory for Computer Science
  • Gilles Kahn, Vice-President for Science, INRIA, The French National Institute for Research in Computer Science and Control, Member of the French Academy of Sciences
  • Prof. Hidehiko Tanaka, University of Tokyo, Japan, Chairman, Promotion Committee of the Real World Computing Project of MITI. (formerly President, Japanese Society for Artificial Intelligence; and Chairman, Promotion Committee for the Fifth Generation Computer Systems project)
  • Prof. Yoichi Muraoka, Waseda University, Japan
  • Prof. Michael K. Simpson, President, American University of Paris
19:00-21:30 Conference Dinner in the Louvre
Friday, October 16
09:00-10:00 Keynote Speaker: Maurice V. Wilkes, Staff Advisor on Research Policy, Olivetti and Oracle Research Lab, Cambridge, UK
CMOS Workstations and Servers---How Far Can Evolution and Innovation Take Us? || Keynote speech in HTML
10:00-10:25 Break
10:25-11:40 Session 11: ILP topics
Chair: Bob Rau, HP Labs
Dynamic Hammock Predication for Non-predicated Instruction Set Architectures , A. Klauser, Univ. of Colorado; T. Austin, Intel; D. Grunwald, Univ. of Colorado; B. Calder, Univ. of California at San Diego
A Direct-Execcution Framework for Fast and Accurate Simulation of Superscalar Processors, V. Krishnan, J. Torrellas, Univ. of Illinois
Efficient Edge Profiling for ILP-Processors A. Eichenberger, S. Lobo, North Carolina State Univ.
11:40-11:55 Break
11:55-13:10 Session 12: Cache locality improvement techniques
Chair: Paul Feautrier, Univ. Versailles St-Quentin, France
A Matrix-Based Approach to the Global Locality Optimization Problem || Presentation slides, M. Kandemir, Syracuse Univ.; A. Choudhary, Northwestern Univ.; J. Ramanujam, Louisiana State Univ.; P. Banerjee, Northwestern Univ.
Transformations for Improving Data Access Locality in Non-perfectly Nested Loops, D. Kulkarni, IBM
Improving Cache Behavior of Dynamically Allocated Data Structures || Presentation slides, D. Truong, F. Bodin, A. Seznec, IRISA/INRIA, France
13:10-14:25 Lunch break
14:25-15:15 Session 13: Data dependence and flow analysis
Chair: Jim Dehnert, SGI
Instance-wise Reaching Definition Analysis for Recursive Programs Using Context-free Transductions || Presentation slides, A. Cohen, J. Collard, Université de Versailles, France
Data Dependence Analysis of Assembly Code, W. Amme, P. Braun, Univ. of Jena, Germany; F. Thomasset, INRIA, Rocquencourt, France; E. Zehendner, Univ. of Jena, Germany
15:15-15:25 Break
15:25-16:20 Session 14: Task scheduling and dynamic load balancing
Chair: Michel Cosnard, Loria, France
Load Balancing in Individual-Based Spatial Applications, M. Dillencourt, L. Bic, F. Merchant, UC Irvine
A New Heuristic for Scheduling Parallel Programs on Multiprocessor, J. Liou, AT&T Laboratories; M. Palis, Rutgers Univ.
Adaptive Scheduling of Computations and Communications on Distributed Memory Systems, M. Al-Mouhamed, H. Najjari, King Fahd Univ., Saudi Arabia
16:20-16:35 Break
16:35-17:35 Session 15: Topics in compilation
Chair: Christian Lengauer, Univ. Passau, Germany
Using Algebraic Transformations to Optimize Expression Evaluation in Scientific Code, J. Zory, F. Coelho, École des Mines de Paris, France
Integrated Compilation and Scalability Analysis for Parallel Systems || Presentation slides, C. Mendes, National Institute of Space Research, Brazil; D. Reed, Univ. of Illinois
Improving Compiler and Run-Time Support for Adaptive Irregular Codes, H. Han, C. Tseng, Univ. of Maryland
On the parallelization of Benchmarks for SSM Machines Y. Paek, New Jersey Institute of Technology; A. Navarro, E. Zapata, Univ. of Malaga, Spain; D. Padua, Univ. of Illinois
17:35-17:50 Break
17:50-18:35 Session 16: Compilation for distributed memory
Chair: Makoto Amamiya, Kyushu Univ., Japan
Efficient Methods for Multi-Dimensional Array Redistribution Y. Chung, C. Hsu, Feng Chia Univ., R.O.C.
A multithreaded runtime environment with thread migration for a HPF data-parallel compiler || Presentation slides, L. Bougé, ENS Lyon, France; P. Hatcher, Univ. New Hampshire; R. Namyst, C. Perez, ENS Lyon, France
Optimized Code Generation for Heterogeneous Computing Environment using Parallelizing Compiler TINPAR, S. Goto, Sumitomo Electric Industries,Ltd.; A. Kubota, Hiroshima City Univ.; T. Tanaka, M. Goshima, S. Mori, Kyoto Univ., Japan; H. Nakashima, Toyohashi Univ. of Tenology; S. Tomita, Kyoto Univ., Japan
18:35-18:40Best paper award presentations, and closing remarks
Saturday, October 17: Tutorials
09:00-12:00 Trimaran Compiler Infrastructure, B. Goldberg, H. Kim, K. Palem, NYU; B. Rau, HP Labs
14:00-17:00 The Tera Multithreaded Architecture, J. Feo, P. Briggs, Tera
Monday, October 19
09:00-17:00 Workshop on Tera Programming

Back to PACT'98 homepage