Abstract:We have developed a general-purpose and high performance parallel visualization library in the HEC-MW project. Three parallel frameworks are provided in our library to satisfy different requirements of applications including a concurrent-in-memory style for extremely large data size, a concurrent-in-time (MxN) style for computational steering and Grid environment, and a server-client style for interactive visualization. Meanwhile, it can support most typical grid types widely used in numerical simulations including regular grids for FDM (Finite Difference Method), particles for DEM (Discrete Element Method), and unstructured grids for FEM (Finite Element Method). Moreover, high speedup performance has been achieved by some hardware-oriented optimization strategies. Different algorithms, data structures and parallelization techniques were adopted for different platforms from PC-clusters to the Earth Simulator. Good visualization images and high speedup performance have been achieved for some applications with different grid types on typical parallel platforms, thus demonstrating the feasibility and effectiveness of our library.1) http://www.rss21.iis.u-tokyo.ac.jp/2) Earth Simulator Research and Development Center web site http://www.es.jamstec.go.jp/3) http://www.kgt.co.jp/4) Lacroute, P. and Levoy, M.: Fast Volume Rendering Using a Shear-Warp Factorization of the Viewing Transformation, ACM SIGGRAPH 94 Proceedings (1994) pp. 451-457.5) Sano, K., Kitajima, H., et al..: Parallel Pocessing of the Shear-Warp Factorization with the Binary-swap Method on a Distributed-memory Multiprocessor System, Proceedings of the IEEE symposium on Parallel rendering, Phoenix, United States (1997) pp.78-94.6) Chen, L., Fujsihiro, I., et al.: Detail-Preserving Resampling for Parallel Volume Rendering Large Unstructured Data on Earth Simulator, Proceedings of The 7th IASTED Conference on Computer Graphics and Imaging, Hawaii (2004) pp. 242-2477) Lorensen, W. E. and Cline, H. E.: Marching Cubes: A High Resolution 3D Surface Construction Algorithm, ACM Comp. Graph, Vol.21, No.4 (1987) pp. 163-169.

A patch-based data reorganization method for coupling large-scale simulations and parallel visualization

A parallel computing method for irregular work

Efficient Coupling of Parallel Visualization and Simulations on Tens of Thousands of Cores

A Parallel Simulator for Massive Reservoir Models Utilizing Distributed-Memory Parallel Systems

Large Scale Numerical Simulation Via Parallelization and Reconfigurable Computing Hardware

Parallel Visualization for Large-Scale Datasets

Parallel visualization of large-scale multifield scientific data

A Virtual Plant Ecosystem Featuring Parallel Computing and Distributed Visualization

Multiresolution Organization of Large-Scale Simulation Data

Parallel Volume Rendering Based on LOD Method

CLUSTER-BASED OCEAN REMOTE SENSING IMAGE FUSION PARALLEL COMPUTING STRATEGY

Optimizing Parallel Performance of Streamline Visualization for Large Distributed Flow Datasets

A Mechanism of Remote Sensing Image for Parallel Processing Base on Splitting Blocks

Correlation Statistical Modeling Reduction Method for Large-Scale Structural Grid Data

Advection-Based Sparse Data Management For Visualizing Unsteady Flow

A108 Parallel Visualization Library for Large-Scale Datasets on Various Types of Platforms from PC-Clusters to Earth Simulator

Large-Scale Datasets Visualization on Networks Using Dynamic Pipeline Decomposition and Mapping

A Parallel Approach for the Generation of Unstructured Meshes with Billions of Elements on Distributed-Memory Supercomputers

Parallel Visualization Based on Heterogeneous Computation Resources

Visualization of Large Non-Trivially Partitioned Unstructured Data with Native Distribution on High-Performance Computing Systems

Large-Scale Parallel Visualization on Earth Simulator