Scientific Data Management Challenges in High-Performance Visual Data Analysis - Scientific Data Management - page 325

Chapter 9

Scientific Data Management
Challenges in High-Performance
Visual Data Analysis

E. Wes Bethel, 1 Prabhat, 1 Hank Childs, 2 Ajith Mascarenhas, 3
and Valerio Pascucci 4
1 High Performance Computing Research Department, Lawrence Berkeley
National Laboratory, Berkeley, California
2 Computing Applications and Research, Lawrence Livermore National
Laboratory, Livermore, California
3 Center for Applied Scientific Computing, Lawrence Livermore National
Laboratory, Livermore, California
4 Scientific Computing and Imaging (SCI) Institute, School of Computing,
University of Utah, Salt Lake City, Utah

Contents

9.1 Introduction .......................................................... 326

9.2 Production-Level, Parallel Visualization Tool Perspective

on SDM ............................................................... 327

9.2.1 How Data Is Processed ....................................... 328

9.2.1.1 I/O ................................................... 328

9.2.1.2 Processing ............................................ 329

9.2.1.3 Rendering and Remote/Distributed Visualization .. 330

...................

9.2.2 How Metadata Can Enable Optimizations 332

..................................

9.2.3 Data Models and Semantics 334

9.2.4 A Real-World Production Parallel Visualization Tool ....... 335

9.3 Multiresolution Data Layout for Large-Scale Data Analysis ........ 336

9.3.1 Background ................................................... 336

9.3.2 Hierarchical Indexing for Out-of-Core Access to

Multiresolution Data ......................................... 337

9.3.2.1 Hierarchical Subsampling Framework ............... 338

9.3.2.2 Binary Trees and the Lebesgue

Space-Filling Curve .................................. 340

9.3.2.3 Performance .......................................... 344

325