CyberBridges Integrating Advanced Grid Infrastructure with Science and Engineering TeraGrid 07 June

1
CyberBridgesIntegrating Advanced Grid
Infrastructure with Science and Engineering
TeraGrid 07June 4-8, 2007 . Madison,
WIEducation and Outreach SessionJune 6th 930
1000 AM

• Presented by Heidi L. Alvarez, Ph.D.
• Director, Center for Internet Augmented Research
  and Assessment

2
CyberBridges

• CyberBridges tested the hypothesis that students
  can act as
• bridges, showing researchers how they can use
• cyberinfrastructure (CI) to further their
  knowledge.
• CyberBridges NSF Award OCI-0537464 to Florida
  International University Center for
• Internet Augmented Research Assessment

TeraGrid 07 paper authors are the Co-PIs,
faculty advisors, and Ph.D. student fellows of
the CyberBridges Pilot Project Heidi L. Alvarez,
Julio Ibarra, Kuldeep Kumar, Chi Zhang, Donald A.
Cox, Eric Johnson, David Chatfield, Eric
Crumpler, Cassian DCunha, Ronald Gutierrez, Tom
Milledge, Giri Narasimhan, Rajamani S. Narayanan,
Alejandro de la Puente, S. Masoud Sadjadi
3
CIARA Background for the Project

• The Center for Internet Augmented Research and
  Assessment (CIARA) was created in 2003 as a State
  of Florida Type II Research Center at FIU. CIARA
  services institutional collaborators in the U.S.
  and internationally as a bridge linking
  researchers and educators with the infrastructure
  and knowledge they need to perform their work

4
AMPATH

• Improves research and education high performance
  network connectivity in the Western Hemisphere
• Promotes efficient peering through a distributed
  exchange point model
• Leverages participants network resources to
  foster collaborative research and advance
  education throughout the Western Hemisphere, to
  Europe and Asia
• AMPATH provides CyberBridges fellows with
  connectivity to RE computational grid resources
  at institutions and laboratories around the world

5
Western HemisphereResearch EducationNetwork
(WHREN) Links Interconnecting Latin America (LILA)

• (1) 2.5Gbps dark fiber segment
• U.S. landings in Miami and San Diego
• Latin America landings in Sao Paulo and Tijuana
• 5M over 5 years from the National Science
  Foundation 5M from the Science Foundation of
  Sao Paulo
• OCI Award 0441095

6
What is the CyberBridges Pilot Project?

• Funded by the NSF Cyberinfrastructure
  (CI)-Training, Education, Advancement, and
  Mentoring (TEAM) program solicitation in 2005
• 11 Awards were made for 250,000 each
• CI-TEAM 2005 was a demonstration program
• Expanded in 2006 to a multi-year implementation
  program
• CI-TEAM increases the effectiveness,
  penetration, and utilization of CI

http//www.nsf.gov/pubs/2005/nsf05560/nsf05560.htm

7
(No Transcript)
8
Center for High Energy Physics Research and
Educational Outreach (CHEPREO) Tier3 Cluster

• Current Cluster configuration
• Gigabit Copper backbone with GigE connectivity to
  AMPATH Network
• OSG Computing Element
• 23 node computational cluster
• Supports local user logins
• Serves as the CEs gatekeeper
• Plus all other cluster services
• Augments CyberBridges cluster
• Funded by NSF MPS-0312038.
• www.chepreo.org

9
CyberBridges Fellows

• In 2005-06 CyberBridges funded four Ph.D.
  students from
• Physics
• Biochemistry
• Bioinformatics / Computer Science
• Biomedical Engineering
• CyberBridges has helped these students and their
  faculty
• advisors transform their research by connecting
  them with CI.

10
Approaches to e-Learning Sciences
CyberBridges fellows construct with their faculty
advisors models and reproduce these connections
through computer simulations. Technology
permits teams, possibly in different locations
and with different skills to work toward solution
of the real work
11
Project structure
4 Fellowships
-Two semesters long -Interdisciplinary
Collaboration -Annual Assessment
-Normative analysis -Case study package
Certification Program
12
Hypothesis and assessment

• CyberBridges proposed that graduate students
  engaged in inquiry-based learning activities can
  affect transfer of CI research and that this
  transfer will increase scientists rates of
  discovery and create a CI-empowered workforce.

Faculty Student CI-Scientist
Domain specific metrics
CyberBridges Assessment
13
Project implementation

• 1. Continuous cross-disciplinary dialogue
• 2 .Collaborative proposal writing
• 3. Faculty review of proposals
• 4. Normative assessment
• 5. Nesting of graduate students in lab
• 6. CI scholastic certification program
• 7. Case study package
• 8. Student outreach

14
Case Studies

• 1Alejandro de la Puente (student) and Rajamani
  S. Narayanan (faculty)
• Interplay between Random Matrix Theory and
  Quantum Field Theory
• Physics
• 2Cassian DCunha (student) and David Chatfield
  (faculty)
• Parallel Computations for Macromolecular
  Simulation
• Chemistry and Biochemistry
• 3Tom Milledge (student) and Giri Narasimhan
  (faculty)
• Unsupervised Pattern Discovery in Protein
  Structures
• Computing and Information Sciences
• 4Ronald A. Gutierrez (student) and Eric T.
  Crumpler (faculty)
• Computational Enhanced Mesh Design in Tissue
  Engineering Measuring Wall Shear Stress in
  Cell/Scaffolds
• Biomedical Engineering

15
Interplay between Random Matrix Theory and
Quantum Field Theory- Alejandro de la Puente

• What is the project trying to do?
• Lattice Field Calculation in Quantum
  Chromodynamics
• Why is it important?
• Enables the study of Quantum Chromodynamics in a
  nonperturbative way.
• What is the expected output?
• Calculations using the lattice formulation will
  enable us to study many physical observables of
  the strong interaction, in particular, the study
  of phase transitions.
• Case Study
• Why study Lattice Field Theories
• How to go from the continuum to the lattice such
  that calculations can be implemented into a
  computer algorithm

16
Project Details

• Determine the fundamental parameters that
  describe each particle in the current standard
  model, to exceed traditional Quantum Field
  Theory.
• Quantum Chromodynamics (QCD) is a theory that has
  proven to be very accurate in the determination
  of these parameters.
• Lattice Quantum Chromodynamics (LQCD), which is a
  discretized version of QCD, replaces the
  continuum with a four-dimensional grid for its
  calculations.
• This is a more complete, accurate (as well as
  more computationally intensive) approach to
  identify particles in data that will be generated
  by the CERN Large Hadron Collider (LHC)
  experiments beginning in 2007.
• Organize and divide up the matrices/computational
  tasks so that Grid computing can be applied.
  Results are compared to the currently well
  accepted LQCD simulation

17
Parallel Computations for Macromolecular
Simulation- Cassian DCunha

• What is the project trying to do?
• Computer simulations of biological macromolecules
• Why is it important?
• To accelerate drug discovery for Pharmaceutical
  industry
• What is the expected output?
• The research is aimed at elucidating the reaction
  mechanism of Chloroperoxidase (CPO) and at
  predicting useful site-specific mutations for
  fine-tuning the enzymes catalytic activity for
  particular substrates of interest.
• Case Study
• Quantum mechanical (QM) calculations for the
  mechanism work
• Molecular mechanical (MM) calculations for the
  mutation work

18
Project Details

• Computer simulations of biological macromolecules
  are widely used in the pharmaceutical industry to
  accelerate drug discovery and in academic
  research for a host of problems from protein
  folding to enzyme mechanism.
• The sizes of the molecular systems and the time
  scales modeled are memory and computation
  intensive. Computer clusters have become popular
  for such applications, either for running many
  simultaneous simulations to obtain good sampling
  statistics or for distributing a single,
  expensive calculation over many computers.
• DCunha has set up a Linux cluster and
  implemented clustering operating system software
  on the basis of instruction received in the
  CyberBridges program. QM and MM software
  (CHARMM, Turbomole and Gaussian) have been
  installed.
• The parallel implementation of CHARMM has been
  benchmarked with a macromolecular simulation.

19
Unsupervised Pattern Discovery in Protein
Structures- Tom Milledge

• What is the project trying to do?
• To implement unsupervised pattern discovery tools
  for protein structure data by using the
  ggeometric hashing technique
• Why is it important?
• To identify common substructures in proteins
• To create multiple 3-D structural alignments
• To identify functional regions in proteins.
• What is the expected output?
• A database of protein structure patterns.
• Case Study
• Zinc finger domains
• Dehydrogenase domains

20
Project Details

• Pattern Discovery refers to the task of
  identifying relevant and significant
  commonalities or essential differences in related
  data. In bioinformatics applications, patterns
  come in different flavors ??in sequences,
  structures, shapes, images, and in quantitative
  and temporal data.
• A database of proteins and sequence information
  is searched to look for protein structures that
  appear to share potentially useful functions. 3-D
  visualization best reveals these functions.
• Milledge has conducted a number of experiments
  using MPI-based message passing search
  strategies, as well as another method using the
  global shared memory GA Toolkit
  (http//www.emsl.pnl.gov/ docs/global/).
• He has used the implementation of Global Arrays
  on the Blue Gene development machines in IBM
  Rochester, where he was an intern over the summer
  2006.

21
Computational Enhanced Mesh Design in Tissue
Engineering-Ronald Gutierrez

• What is the project trying to do?
• Enable greater computer power to simulate more
  realistic models for bioengineering applications
• Why is it important?
• High Performance Computer (HPC) resources can
  overcome efficiently present obstacles in
  theoretical and experiment observations.
• What is the expected output?
• Three Dimensional (3D) flow field calculation and
  visualization via computational fluid dynamics
• Case Study
• Computational Fluids Dynamics (CFD) on scaffold
  design In Tissue Engineering (TE)
• Taylor vortices visualization present neonatal
  aorta Coarctation

22
Model Parallelization

• CyberBridges Cluster consisted of 6 nodes
  (siblings) Intel Processors 3GHz, OSRocks
• Grid Convergence and Memory requirements
• 1.6 Million unstructured cells 2.2 Gb Ram for
  single processor (Windows and Linux Machines
  cannot address more than 2 Gb Ram)
• Computational Simulation Results after 3 hours
  (1 Module-Flow only) using 4 nodes.
• Results of TE scaffolds models contributed to the
  FLOW FORCES distribution analysis and hence an
  improved design was accepted
• 3D visualization offered undeniable advantages
  over 2D to evaluate geometrical characteristics
  and hydrodynamic performance of the TE scaffolds

23
3D CFD Models Advantages

• Realistic Modeling -vs.-Time consuming
  experiments
• Ability to simulate both realistic and ideal
  conditions
• Prediction and control of equipment
• Cost effective
• Measurements for inaccessible situations can be
  overcome

3D Disc and Oblong streamlines ( Source
Gutierrez R., Crumpler E., submitted Publication
2006 Annals of Biomedical Engineering)
24
Outcomes of CyberBridges Pilot

• Re-focusing expenditures from technology
• support functions to synchronous fellowships
• can have a profound impact on the
• effectiveness of scientific investigations.
• CyberBridges fellows and their advisors have
• benefited from such re-focusing. They have
• successfully implemented CI into their
• research and teaching.
• Globally distributed institutions can replicate
• the CyberBridges model.

25
CyberBridges Success

• CyberBridges exposed over 200 students
• and researchers to CI during SC06, when
• the four CyberBridges fellows gave talks
• on their work.
• CyberBridges made it possible for each
• of the fellows to acquire competitive
• summer internships.

26
CyberBridges Fellows Present Graduate
SuperComputing 2006, Tampa, Florida November
11-17 www.sc06.supercomp.org
Top from left Professor Yan BaoPing, Chinese
Academy of Sciences Professor Paul Avery, UF
Dr. Miriam Heller, NSF Tom Milledge Ronald
Gutierrez Alejandro de la Puenta Cassian
DCunha Bottom row from left Ernesto Rubi,
FIU/CIARA Michael Smith, FIU/CIARA Dr. Eric
Crumpler, FIU Engineering Julio Ibarra, Co-PI
Executive Director CIARA Heidi Alvarez, PI
Director CIARA, Dr. S. Masoud Sadjadi, FIU SCIS.
27
Presentation Publication

• Ronald Gutierrez presenting CyberBridges A
  Model Collaboration Infrastructure for
  e-Science", Proceedings of the 7th IEEE
  International Symposium on Cluster Computing and
  the Grid (CCGrid), Rio de Janeiro, Brazil, May
  2007 at CCGRID
• After the presentation, I met some PhD students
  that wanted to know if I could present this
  collaboration program (CyberBridges) in another
  conference (CLCAR-07, Santa Marta, Colombia).

28
NSF CI-TEAM Implementation Project Global
CyberBridges (GCB) A Model Global Collaboration
Infrastructure for e-Science between US and
International Partners

• Trans-national and cross-discipline communication
  is the future for
• science and engineering research and education.
• Global CyberBridges extends the CyberBridges
  concept to an
• international level.
• International Partners
• Florida International University
• Computer Network Information Center of the
  Chinese Academy of Sciences
• City University of Hong Kong
• University of Sao Paulos School of the Future in
  Brazil.

29
Questions?
LambdaVision 100-Megapixel display and SAGE
(Scalable Adaptive Graphics Environment) software
developed by the Electronic Visualization
Laboratory at the University of Illinois at
Chicago. Major funding provided by NSF.
Email info_at_cyberbridges.net Website
www.cyberbridges.net