Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Accelerating In-transit Isosurface Generation With Topology Preserving Compression

Yanliang Li, Jieyang Chen

TL;DR

The paper tackles the data-transfer bottleneck in in-transit isosurface visualization by introducing a topology-preserving, lossy compression framework. It builds a Compressed Hierarchical Representation ($CHR$) through domain decomposition, isovalue-based data pruning, and block-level indexing over candidate isovalues, then applies topology-preserving compression to bound isosurface error by translating it to an allowable raw-data error, with per-edge distances such as $d_0$ and $d_1$ used to form the bound. The approach yields end-to-end speedups up to $2.6\times$ for exact topology preservation and up to $4\times$ when permitting topology accuracy deviations (e.g., $99\%$, $95\%$, $80\%$). This methodology enables near real-time, in-transit isosurface visualization for large-scale scientific workflows, with broad applicability to CFD, medical imaging, and geophysics.

Abstract

Data visualization through isosurface generation is critical in various scientific fields, including computational fluid dynamics, medical imaging, and geophysics. However, the high cost of data sharing between simulation sources and visualization resources poses a significant challenge. This paper introduces a novel framework that leverages lossy compression to accelerate in-transit isosurface generation. Our approach involves a Compressed Hierarchical Representation (CHR) and topology-preserving compression to ensure the fidelity of the isosurface generation. Experimental evaluations demonstrate that our framework can achieve up to 4x speedup in visualization workflows, making it a promising solution for real-time scientific data analysis.

Accelerating In-transit Isosurface Generation With Topology Preserving Compression

TL;DR

The paper tackles the data-transfer bottleneck in in-transit isosurface visualization by introducing a topology-preserving, lossy compression framework. It builds a Compressed Hierarchical Representation () through domain decomposition, isovalue-based data pruning, and block-level indexing over candidate isovalues, then applies topology-preserving compression to bound isosurface error by translating it to an allowable raw-data error, with per-edge distances such as and used to form the bound. The approach yields end-to-end speedups up to for exact topology preservation and up to when permitting topology accuracy deviations (e.g., , , ). This methodology enables near real-time, in-transit isosurface visualization for large-scale scientific workflows, with broad applicability to CFD, medical imaging, and geophysics.

Abstract

Data visualization through isosurface generation is critical in various scientific fields, including computational fluid dynamics, medical imaging, and geophysics. However, the high cost of data sharing between simulation sources and visualization resources poses a significant challenge. This paper introduces a novel framework that leverages lossy compression to accelerate in-transit isosurface generation. Our approach involves a Compressed Hierarchical Representation (CHR) and topology-preserving compression to ensure the fidelity of the isosurface generation. Experimental evaluations demonstrate that our framework can achieve up to 4x speedup in visualization workflows, making it a promising solution for real-time scientific data analysis.
Paper Structure (6 sections, 4 figures)

This paper contains 6 sections, 4 figures.

Table of Contents

  1. Introduction
  2. Compression Accelerated In-transit Isosurface Generation
  3. Compressed Hierarchical Representation
  4. Topology Preserving Compression
  5. Experimental Evaluation
  6. Conclusion

Figures (4)

  • Figure 1: Our accelerated in-transit isosurface generation framework
  • Figure 2: Compressed Hierarchical Representation (CHR) generation process
  • Figure 3: Isosurface generation with and without topology preservation
  • Figure 4: End-to-end time breakdown of in-transit isosurface generation with and without our proposed topology preservation compression. Two block sizes ($64^3$ and $128^3$) for domain decomposition are tested.