Languages and Compilers for High Performance Computing: 17th International Workshop, LCPC 2004, West Lafayette, IN, USA, September 22-24, 2004, Revised Selected Papers

ISBN-13: 9783540280095 / Angielski / Miękka / 2005 / 494 str.

Rudolf Eigenmann;Zhiyuan Li;Samuel P. Midkiff

Opis

This book constitutes the thoroughly refereed post-proceedings of the 17th International Workshop on Languages and Compilers for High Performance Computing, LCPC 2004, held in West Lafayette, IN, USA in September 2004.

The 33 revised full papers presented were carefully selected during two rounds of reviewing and improvement. The papers are organized in topical sections on compiler infrastructures; predicting and reducing memory access; locality, tiling, and partitioning; tools and techniques for parallelism and locality; Java for high-performance computing; high-level languages and optimizations; large-scale data sharing; performance studies; program analysis; and exploiting architectural features.

Spis treści

Experiences in Using Cetus for Source-to-Source Transformations.- The LLVM Compiler Framework and Infrastructure Tutorial.- An Overview of the Open Research Compiler.- Trimaran: An Infrastructure for Research in Instruction-Level Parallelism.- Phase-Based Miss Rate Prediction Across Program Inputs.- Speculative Subword Register Allocation in Embedded Processors.- Empirical Performance-Model Driven Data Layout Optimization.- Implementation of Parallel Numerical Algorithms Using Hierarchically Tiled Arrays.- A Geometric Approach for Partitioning N-Dimensional Non-rectangular Iteration Spaces.- JuliusC: A Practical Approach for the Analysis of Divide-and-Conquer Algorithms.- Exploiting Parallelism in Memory Operations for Code Optimization.- An ILP-Based Approach to Locality Optimization.- A Code Isolator: Isolating Code Fragments from Large Programs.- The Use of Traces for Inlining in Java Programs.- A Practical MHP Information Analysis for Concurrent Java Programs.- Efficient Computation of Communicator Variables for Programs with Unstructured Parallelism.- Compiling High-Level Languages for Vector Architectures.- HiLO: High Level Optimization of FFTs.- Applying Loop Optimizations to Object-Oriented Abstractions Through General Classification of Array Semantics.- MSA: Multiphase Specifically Shared Arrays.- Supporting SQL-3 Aggregations on Grid-Based Data Repositories.- Supporting XML Based High-Level Abstractions on HDF5 Datasets: A Case Study in Automatic Data Virtualization.- Performance of OSCAR Multigrain Parallelizing Compiler on SMP Servers.- Experiences with Co-array Fortran on Hardware Shared Memory Platforms.- Experiments with Auto-Parallelizing SPEC2000FP Benchmarks.- An Offline Approach for Whole-Program Paths Analysis Using Suffix Arrays.- Automatic Parallelization Using the Value Evolution Graph.- A New Dependence Test Based on Shape Analysis for Pointer-Based Codes.- Partial Value Number Redundancy Elimination.- Overflow Controlled SIMD Arithmetic.- Branch Strategies to Optimize Decision Trees for Wide-Issue Architectures.- Extending the Applicability of Scalar Replacement to Multiple Induction Variables.- Power-Aware Scheduling for Parallel Security Processors with Analytical Models.