Visualization Streams for Ultimate Scalability

A Research Project in Scientific Visualization centered on the development of cache oblivious approaches for the management, streaming and rendering of large surface and volume meshes.
Principal Investigator: Valerio Pascucci
Other team members: Eric Lamar, Daniel Laney, Peter Lindstrom.
Main Collaborators: Mark Duchaneau, Randall J. Frank, Giorgio Scorselli, Dave Bremer.
Students:  Janine Bennett (Ph.D.), Rita Borgo (Ph.D.), Peer-Timo Bremer (Ph.D.), Kree Cole-McLaughlin (BS), Ajith Marscarenas (Ph.D.), Vijay Natarajan (Ph.D.), Sinesio Pesco (Postdoc), Jonathan Strasser (MS).

In the ViSUS project (see the featured article in the LDRD report) we develop data streaming techniques for progressive processing and visualization of large scientific datasets. Our strategy is to exploit the coupling between time-critical algorithms and progressive multi-resolution data-structures to realize an end-to-end optimized flow of data from the original source, such as remote storage or large scientific simulation, to the rendering hardware.
The implementation of this approach will enable three major visualization modalities. (i) Interactive visualization on high resolution power-walls. (ii) Interactive visualization on desktop workstations of large datasets that cannot be stored locally. (iii) Immediate monitoring of remote simulations from a desktop workstation.
These modalities target multiple phases in the process of generating and exploring very large simulation datasets where real-time user interaction can increase the productivity of scientists.

Streaming live at Supercomputing 2004!

ViSUS Progressive Viewer (monitor in front) showing in real-time a Miranda simulation executed on BlueGene/L (in the background on the left).

The ViSUS streaming infrastructure enabled the two live demonstrations of LLNL simulation codes (Miranda and Raptor) visualized in real-time while executed on the BlueGene/L prototype installed at the IBM booth of the SC04 exhibit. The same day BlueGene/L was announced to be the new fastest supercomputer in the world.
  On BlueGene/L: the Miranda and the Raptor simulations are executed on the IBM BG/L prototype (in the picture on the left) with a simple set-up that allows saving files in IDX format with the ViSUS streaming library.
   On the linux front-end: the ViSUS data server is activated on the linux front-end to load the simulation data as it appears on disk. A network connection is opened to allow connections from remote visualization applications requesting the data.
   On the graphics workstation: the ViSUS Progressive Viewer is executed on a Windows graphics workstation. At startup, the Progressive Viewer opens a network connection to the Data Server and start streaming the simulation data for visualization. The IDX multi-resolution data model allows adjusting the quality of the visualization depending on the communication speed and on the performance of the local workstation.


ViSUS related publications

Real-Time Monitoring of Large Scientific Simulations
V. Pascucci, D. E. Laney, R. J. Frank, G. Scorzelli, L. Linsen, B. Hamann, F. Gygi,
In Proceedings of the 18-th annual ACM Symposium on Applied Computing, March, 2003, Melbourne, FL, pages 194-198.
(pdf-paper) (pdf-presentation)

Hierarchical Indexing for Out-of-Core Access to Multi-Resolution Data
V. Pascucci, and , R. J. Frank,
Chapter in Hierarchical and Geometrical Methods in Scientific Visualization, 2002, pages 225-241.

Global Static Indexing for Real-time Exploration of Very Large Regular Grids.
V. Pascucci and R. J. Frank,
In proceedings of Super Computing 2001, on-line proceeding, November, 2001.
(pdf-paper) (pdf-presentation)

Implicit Occluders
S. Pesco, P. Lindstrom, V, Pascucci and C. Silva,
IEEE/SIGGRAPH Symposium on Volume Visualization, 2004, pages 47-54.

Encoding Volumetric Grids For Streaming Isosurface Extraction
A. Mascarenhas, M. Isenburg, V. Pascucci, and J. Snoeyink,
In Proceeding of the 2-nd International Symposium on 3D Data Processing, Visualization and Transmission (3DPVT), 2004, to appear.

Progressive Dimension-Independent Boolean Operations
A. Paoluzzi, V. Pascucci, and G. Scorzelli,
In Proceeding of the 9-th ACM Symposium on Solid Modeling and Applications (SM04), 2004, pages 203-211.
(pdf-paper) (mechanical-piece-avi-animation) (temple-avi-animation)

Isosurface Computation Made Simple: Hardware Acceleration, Adaptive Refinement and Tetrahedral Stripping
V. Pascucci,
Joint Eurographics - IEEE TVCG Symposium on Visualization (VisSym), 2004, pages 293-300.

A Multi-Layered Image Cache for Scientific Visualization
E. LaMar, and V. Pascucci,
IEEE Symposium on Parallel and Large-DataVisualization and Graphics (PVG), 2003, pages 61-68.
(pdf-paper) (pdf-presentation)

Using Graphs for Fast Error Term Approximation of Time-varying Datasets
C. Nuber, E. C. LaMar, V. Pascucci, B. Hamann, K. I. Joy,
In Proceedings of Eurographics VisSym 2003, to appear.

InteractiveView-dependent Rendering of Large Isosurfaces
B. Gregorski, M. Duchaineau, P. LIndstrom, V. Pascucci, and , K. I. Joy,
In Proceedings of IEEE Conference on Visualization, 2002, pages 475-482.

Slow Growing Subdivision (SGS) in Any Dimension: Towards Removing the Curse of Dimensionality,
Valerio Pascucci,
Computer Graphics Forum, vol. 21, n. 3, September 2002, pages 451-460, (Proceeding of Eurographics 2002).
(pdf-paper) (pdf-presentation

Terrain Simplification Simplified: A General Framework for View-Dependent Out-of-Core Visualization,
Peter Lindstrom and Valerio Pascucci,
IEEE Transactions on Visualization and Computer Graphics, vol. 8, n.3, July-September 2002, pages 239-254.
(pdf-paper) (mpeg-animation)

Temporal and Spatial Level of Details for Dynamic Meshes.
A. Shamir and V. Pascucci,
In proceedings of ACM Symposium on Virtual Reality Software and Technology, pages 77-84, November, 2001.
(pdf-paper) (pdf-presentation)

Visualization of Large Terrains Made Easy.
P. Lindstrom and V. Pascucci,
In proceedings of IEEE Conference on Visualization, 2001, pages 363-370,574.
(pdf-paper) (pdf-presentation) (source-code, data, addendum, maintained by Peter)

Time critical isosurface refinement and smoothing,
V. Pascucci, and C. L. Bajaj,
In Proceeding of the ACM Symposium on Volume Visualization and Graphics (VolVis), 2000, pages 22-42.
(pdf-paper) (pdf-presentation)

Multi-Resolution Dynamic Meshes with Arbitrary Deformations.
A. Shamir, V. Pascucci, and C. L. Bajaj,
In Proceedings of the IEEE Conference on Visualization, 2000, pages 423-430.
(pdf-paper) (pdf-presentation)

Parallel accelerated isocontouring for out-of-core visualization.
C. L. Bajaj, V. Pascucci, D. Thompson, and X. Y. Zhang,
Proceedings of the 1999 IEEE symposium on Parallel visualization and graphics (PVG), 1999, pages 97-104.