06/13/00 Visualization, Analysis, and Imaging Lab (VAIL) Capabilities Robert Moorhead, Director Mississippi State University

06/13/00 Overview of Slides Present Products (in various states of further development) –EnVis (batch mode rendering) –ISTV (interactive visualization) –cthru (virtual environment for ocean models) Future Products (in various states of development) –focusing on an integrated system, rather than a set of products –DataManager (to ingest datasets) –dmvis (small toolkit to test other parts of system) –Triton (new virtual environment for littoral region)

06/13/00 EnVis batch model rendering system use EnVis to produce images use HUM (open source program) to interactively pan and zoom while looping through sequence of images 1/32  degree global Navy Layered Ocean Model (NLOM) produces 8192x4608 pixel image when mapped data point to pixel. unable to see that resolution on most workstations –zoom in? –examine subregion? takes about 15 sec/image to render 8192x4608 image

06/13/00 EnVis

06/13/00 ISTV Interactive Structured Time-varying Visualizer (ISTV) designed specifically for large scale ocean models developed in collaboration with NRL/SSC monolithic but fast installed at NRL/SCC, NAVO/MSRC, and ERDC/MSRC been used at NRL/SSC for oceanography been used at NAVO for hydrographic data analysis been used at CEWES for WAM and CH3D studies runs on Irix and Solaris optimized for large data vis (dynamic ocean models) application-specific readers and calculator functions

06/13/00 ISTV

06/13/00 cthru VE for studying dynamic physical oceanography data developed with NRL/SSC and NAVO, with much assistance from NCSA 4 processes: –master process –remotely-executing ocean model –extraction module –visualization module has been used to study Sea of Japan (NLOM), Chesapeake Bay (NCOM), and Pacific Ocean (NLOM)

06/13/00 cthru

06/13/00 Execution Manager Integrated System Diagram Extraction Manager Data Manager Graphics Server disk

06/13/00 Integrated System Focus One Data Manager (DM) for all visualization paradigms DM runs in close to data source Use DM for –batch (e.g., EnVis) –interactive (dmvis) vtk (for portability and functionality) custom vis tools (for performance) –virtual environment (Triton)

06/13/00 Remote Visualization Distributed Visualization Environment –model/extraction on one machine –graphics on a separate machine –integrate/use data manager –models: ADCIRC, NCOM, COAMPS, ??

06/13/00 Remote Visualization Status –designed and implemented a DM server/client a prototype for the full-scale DataManager a foundation on which to build a distributed VE –current implementation of DM includes multiple socket connections to talk to vis front-end multi-threading

06/13/00 Remote Visualization Status (cont) –server allows queries and requests for data from multiple applications running concurrently –client currently includes a crude caching scheme and non-blocking requests for data from an application –vis (dmvis) routines currently include arbitrary cutting planes, isosurfaces, and bathymetry

06/13/00 Remote Visualization dmvis –core functionality is the same for Interactive Virtual Environment (VE) –VR Juggler added to create VE -- Triton –VR Juggler (work-in-progress) uses pthreads, which are more standard than CAVELib’s sprocs –compiles and runs on Solaris (Sun) and Irix (SGI)

06/13/00 Why not extend cthru? cthru was designed to ingest output from a single ocean model cthru was designed to ingest data on a static structured grid in latitude/longitude (but time-varying in z, e.g., NLOM) cthru was designed for concurrent visualization of running ocean models cthru was built using circa 1995 technology and design principles (it’s not object-oriented) need a VE that can ingest multiple models from same 3D space that have different spatio-temporal gridding

06/13/00 Future Work optimized performance of distributed VE integration of data manager and VE ultimate goal is for data manager to support –visualization in batch, interactive, and virtual environments –visualization of multiple simulation model outputs (NCOM, ADCIRC, etc.) and associated measured/sensed data (bathymetry, wind fields, coastlines, etc.) in same view volume.