Title: TeraGrid Visualization Tutorial
1TeraGrid Visualization Tutorial
- TeraGrid 2007
- June 4, 2007
2Tutorial Schedule (Morning)
- 830 900 Introduction and Open Discussion
- What is the purpose of the tutorial?
- What are attendees hoping to gain from the
tutorial? - Have you tried using TG vis resources
previously? - What is good about TG vis resources and what is
missing? - 900 1000 Overview of TG Visualization
Resources - Mission and Vision
- Hardware Resources
- General Visualization Applications that Users
Can Expect - Custom Applications Being Built
- TeraGrid Visualization Gateway
- 1000 1030 Break
- 1030 1115 TeraGrid Visualization Gateway
Hands-On - 1115 1200 Discussion Session about Users
Data - Discuss What Attendees Data Looks Like
- Discuss and Evaluate Potential Applications
3Tutorial Schedule (Afternoon)
- 1200 - 130 Lunch
- 130 300 Interactive Visualization Examples
- EnVision
- Hand On Visualization Examples
- 300 - 330 Break
- 330 - 430 Hands On Session
- 430 - 500 Wrap Up and Discussion
4Introduction and Open Discussion
- What is the purpose of the tutorial?
- What are attendees hoping to gain from the
tutorial? - Have you tried using TG vis resources
previously? - What is good about TG vis resources and what is
missing?
5Overview of TeraGrid Visualization Resources
6TeraGrid Vision
- TeraGrid is a national production
cyberinfrastructure facility whose vision is to
create integrated, persistent, and pioneering
computational resources that will significantly
improve our nations ability and capacity to gain
new insights into our most challenging research
questions and societal problems. - This vision requires an integrated approach to
the entire scientific workflow including
obtaining access, application development and
execution, data analysis, collaboration and data
management.
7TeraGrid Objectives
- DEEP Science Enabling Petascale Science
- Make Science More Productive through an
integrated set of very-high capability resources - Address key challenges prioritized by users
- WIDE Impact Empowering Communities
- Bring TeraGrid capabilities to the broad science
community - Partner with science community leaders - Science
Gateways - OPEN Infrastructure, OPEN Partnership
- Provide a coordinated, general purpose, reliable
set of services and resources - Partner with campuses and facilities
8What Role Does Visualization and Data Analysis
Play in the TeraGrid?
- We focus explicitly on data movement, data
management, data collections and visualization as
value added services that enhance and compliment
the core capabilities of the TeraGrid. - Visualization is managed through the
Visualization Working Group and provides
resources and services to the user community. - Data Movement and data management are managed
through the Data Working Group that deploys
services and tools through the Common TeraGrid
Software Stack (CTSS) - Data Collections are relatively new to the
TeraGrid and are managed through a newly formed
Data Collections Working Group that define the
provision, growth and expansion of data
collections as resources provided by the TeraGrid.
9TeraGrid Visualization Strategy and Mission
- Combine existing resources and current
technology - Commodity clusters and commodity graphics
- Terascale visualization clusters
- Grid technology
- Efforts, expertise, and tools from each of the TG
sites - to enable new and novel ways of visually
interacting with and gaining insight into science
through the analysis of simulations and data. - TeraGrid deploys 2 visualization resources, but
has 9 resource provider sites that do active
visualization for the TeraGrid user community. - TeraGrid also deploys visualization resources for
batch, interactive, and collaborative
visualization.
10TeraGrid Visualization Software Stack (VTSS)
11TeraGrid Areas of Interest and Expertise
- Purdue University (Purdue)
- Remote batch rendering
- Remote interactive visualization
- Hardware-assisted remote rendering
- San Diego Supercomputer Center (SDSC)
- Visualization web services
- Volume rendering
- Remote rendering
- High definition resolution visualization
- Texas Advanced Computing Center (TACC)
- Remote and collaborative visualization
- Large data visualization
- Feature detection
- The University of Chicago / Argonne National
Laboratory (UC/ANL) - Visualization, analysis, and collaboration
services - Remote rendering
- Hardware-assisted volume rendering
- Indiana University (IU)
- Distributed software rendering
- Data repository and visualization integration
- National Center for Atmospheric Research (NCAR)
- Analysis tools
- Multiresolution rendering of large data
- Volume rendering
- Domain specific tools
- National Center for Supercomputing Applications
(NCSA) - High definition resolution visualization
- Volume rendering
- Visualization web services
- Oak Ridge National Laboratory (ORNL)
- Visualization of experimental data collected from
spallation neutron source - Pittsburgh Supercomputing Center (PSC)
- Simulation-time Visualization with hardware
rendering directly tied in to the supercomputer
12TeraGrid Visualization Hardware Resources
- University of Texas at Austin/Texas Advanced
Computing Center (TACC) Visualization Resource - Large, shared memory machine with 512 GB of
shared memory and 16 dual out commodity graphics
cards - Uses VNC as a means for remote/collaborative
visualization - Enables any third party visualization package
that uses OpenGL such as - Paraview
- EnSight
- Amira
- VMD
- University of Chicago/Argonne National Lab
Visualization Resource - Cluster configuration, 96 visualization nodes, 62
compute nodes - Focused on a solution that provides a
client-server visualization mechanism through the
use of Paraview
13UT/TACC Visualization Resource
Sun Microsystems E25K Server 64 Dual Core
UltraSparc IV Processors 512GB Shared Memory
High Performance Fiber Channel Interconnect
High Performance Gigabit Ethernet
Sun Microsystems V890 Login 8 Dual Core
UltraSparc IV Processors 32GB of Shared Memory
14UC/ANL Visualization Resource
15Last Years Future
- Development and deployment of community access
environment to the TeraGrid Visualization
Gateway. (Paper Presentation Tomorrow) - Integration of visualization gateway efforts
into a single unified solution. (Will be
viz.teragrid.org) - Development and deployment of hardware-based
volume rendering service. - Integration of Access Grid client into TeraGrid
Visualization Gateway for collaborative viewing
and interaction. - Recruit visualization community to contribute to
and use the TeraGrid Visualization Gateway. (See
Purdues TeraDRE)
16Next Years Future
- Continued development and deployment of
community access environment to the TeraGrid
Visualization Gateway. (Will be viz.teragrid.org) - Continued development and deployment of
hardware-based volume rendering service. - Continued integration of Access Grid client into
TeraGrid Visualization Gateway for collaborative
viewing and interaction. (Dependent on funding) - Recruit visualization community to contribute to
and use the TeraGrid Visualization Gateway. - Renewed focus and effort of the TeraGrid
visualization working group. - Investigation of visualization resources for
Petascale datasets.
17More TeraGrid Visualization Content
- Tuesday, June 5
- Paper - 330 - 400PM
- Chris Gilpin - Visualizing Electron Microscope
Reconstructions Using TeraGrid Resources - Paper - 400 - 430PM
- Enabling Science Through The TeraGrid
Visualization Gateway -
- Posters - Tuesday Night 630 - 830
- TeraGrid Visualization Overview
- TACC RP Pylon ? Tuesday Evening
- The University of Chicago RP Pylon ? Posters
18Types of Accounts on TeraGrid Resources
19Getting an Account on TeraGrid Resources
- Goto https//pops-submit.ci-partnership.org
- Fill out an application for a DAC allocation
- You will be given 10K SUs on TG machines.
20BreakBack at 1030
21TeraGrid Visualization Gateway
22Discussion Session about Users Data
- Attendees Data
- Discuss and Evaluate Potential Applications
23LUNCHWe will resume at 130
24Interactive Visualization Session
25EnVision
- Google Maps for general purpose scientific
visualizations - Goal drastically simplify the visualization
process via semi-automation of data importation,
rendering algorithm selection - Make visualization accessible to non-experts and
casual users while offering compelling new
capability for expert users - Use remote visualization model to enable new
capability versus repackaging existing
visualization software
26Starting Visualization Jobs on Maverick
27(No Transcript)
28Example Scripts in /usr/local/qsub
29Qsub script to start an xterm
30(No Transcript)
31(No Transcript)
32Turbo VNC http//virtualgl.sourceforge.net
33Before running any job set up your vnc password
by typing vncpasswd at the command line prompt
34(No Transcript)
35(No Transcript)
36(No Transcript)
37(No Transcript)
38(No Transcript)
39Visualizing Electron Microscopy Data
40Electron Microscopy Data Dr. Christopher Gilpin
- Data in raw binary format. Use a script to
create .case file - Create isosurface using Isosurface button (7th
from left) - Data is at fine granularity, so use Reduce
Polygons option before launching isosurface
creation (see next slide)This makes surface more
manageable.
41Electron Microscopy Data Dr. Christopher Gilpin
42Electron Microscopy Data Dr. Christopher Gilpin
- ParaView can open .case file directly
- Create isosurface with Contour button (on top,
9th from left) - Apply a Decimate filter to reduce the amount of
geometry in the isosurface - Apply a Smooth filter to reduce surface
artifacts
43Electron Microscopy Data Dr. Christopher Gilpin
44(No Transcript)
45Visualizing Computational Fluid Dynamics (CFD)
Data
46Running EnSight on maverick
- By default, EnSight runs 8 client and 8 server
threads - To change the default, load module then
- export ENSIGHT8_MAX_THREADS 1..128 (server)
- export ENSIGHT8_MAX_CTHREADS 1..128 (client)
- Executable command vglrun ensight8
(vglrun maps OpenGL to X-windows) - Separate client-server commands
- ensight80.server ltserver optsgt
- ensight80.client ltclient optsgt
47CFD Data Dr. George Karniadakis
- Data originally in many small ParaView files
- Wrote script to aggregate filesLoad time shrank
from hours to seconds (!!!) - Left shows single color (change with Actor
Color button) - Right shows color by Pressure variable(select in
Color by menu)
48CFD Data Dr. George Karniadakis
- Data originally not visible when loaded
- Use View-gt3D View Properties from main menu
- Select Camera tab, click a standard view
button(Left shows X view) (Right shows Y
view) - Select General tab and Use Light Kit for
advanced lighting features - Use File-gtSave Data to save data as EnSight
Case
49CFD Data Dr. George Karniadakis
- Load .case file in EnSight loaded as wireframe
- To create shaded surface
- select part in case menu(upper left corner)
- select shading by pressure(paint can icon on
left) - select full geometry representation(Rubiks cube
icon on left) - use shaded surfaces(sun icon on top)
50CFD Data Dr. George Karniadakis
51CFD Data Dr. George Karniadakis
- Export from ParaView saves only single timestep
- Must save individual timesteps and integrate into
case file for animation support - See EnSight user guide for details