NPACI Thrust Area Visions

1
NPACI Thrust Area Visions

• Mark Ellisman
• NPACI Cross-disciplinary Coordinator
• Stephanie Sides
• NPACI Partner Communications
• 2nd Annual NPACI All-hands Meeting
• January 28, 1999

2
A Vision Is

• A plan to achieve an important scientific goal,
  which requires development and implementation
  of innovative computational infrastructure.

3
Visions for the Thrust Areas Are Critical to
NPACIs Success

• Guide the achievement of long-term goals
• Reinforce linkages with other thrusts, including
  EOT
• Are implemented in shorter-term deliverables by
  which to assess progress

4
Visions Should Define New Capabilities that

• Advance science and provide socially useful
  outcomes
• Focus/integrate NPACI projects and thrust areas
• Push the development and use of NPACI
  infrastructure
• Demonstrate the success of NPACI

5
Vision Template From Broad Goal to Increasing
Levels of Specificity

• Level 1 Overarching vision for thrust area
• Level 2 Collection of projects through which the
  vision will be realized
• Level 3 How each project will contribute

6
Progress to Date

• Process began 9 months ago
• Thrust leaders asked to articulate visions for
  their respective thrusts
• Visions based on template in which lower levels
  become increasingly specific
• Integration with other thrusts critical, esp.
  technology with applications to prove usefulness

7
Progress (cont.)

• Visions were developed and have been reviewed
  twice by the Executive Committee.
• This work taking place simultaneously with work
  to develop software integration plan (see next
  session,Software Roadmap).
• This session next presents visions developed to
  date by each thrust leader and concludes with
  where we go from here.

8
Neuroscience

• Enable new understanding of the brain by linking
  data about macroscopic brain function to its
  molecular and cellular underpinnings.
• Provide infrastructure to enable large and
  multi-scale brain mappingfrom subcellular
  supramolecular details to tissue organization.
• Federate and visualize brain-mapping databases
  consisting of PET, MRI, histological,
  laser-microscopy, and 3-D electron microscopy
  data and enhance tools for traversal of scales
  (PTE, DICE, IE, NS, RWG).
• Link federated databases to compute resources and
  software tools (META, DICE).
• Parallelize, optimize, and port
  computation-intensive codes for reconstruction,
  comparison, or analysis to the NPACI compute
  resources and facilitate use (PTE, IE).
• Provide infrastructure to support modeling of
  neuronal propertiesfrom molecular signaling to
  information processing of neuronal systems.
• Deploy the widely used neuronal simulation codes
  and port them to additional compute platforms
  (Genesis, Neuron) (RWG, NS).
• Port Monte Carlo simulation codes for complex
  molecular and electrical events in synaptic
  transmission and apply to problems not currently
  in reach (NS, PTE).

9
Data-intensive Computing

• Automate information discovery across specific
  data collections.
• Enable scientific data publication by 2001.
• Integrate digital library, data-handling, and
  archival storage systems by 2000.
• Provide a data-handling system by 1999 with a
  uniform interface to all NPACI data resources and
  a common information discovery interface.
• Integrate the data-management infrastructure with
  the metacomputing remote execution environment by
  2001.
• Develop a common security infrastructure across
  metacomputing and data-management systems by
  2000.
• Develop data access and discovery interfaces
  between data-management and metacomputing systems
  by 2000.

10
Earth Systems Science

• Build infrastructure to bridge model scales
  (local, regional, global) and exploit data
  explosion to enable next-generation models of
  climate, weather, hydrology, ecology, and natural
  hazards (earthquakes, volcanoes, floods).
• Expose technology developers to computationally
  challenging ESS problems.
• Bring ESS researchers into contact with Enabling
  Technologies infrastructure.
• Demonstrate applications (two years),
  highlighting NPACIs critical enabling role.
• Integrate OGCM/AGCM/MAS with Legion. Incorporate
  POP model of ocean circulation.
• Integrate with REINAS data-gathering system and
  coastal model (San Diego Bay).
• Integrate modeling efforts with EOT.
• Make data collections and modeling results
  accessible on the Web.
• Support new ESS initiatives where NPACI-developed
  infrastructure may play enabling role.
• Pursue KDI solicitation, Plate Boundary
  Observatory initiative, and UC Remote Sensing
  initiative.
• Explore global applications of REINAS.
• General Earthquake Model (simulations from
  microscale to macroscale).
• Workshops Improve NPACI teamwork, identify new
  constituencies, and achieve theoretical advances.
• Develop metrics to gauge success and guide
  evolution and growth.

11
Metasystems

• Provide a distributed, transparent, secure,
  high-performance, collaborative computational
  environment for scientists, engineers, educators,
  and humanists.
• When?
• Nowready in pre-production mode at a number of
  sites.
• In FY 99, Globus and Legion will move into
  pre-production mode (stage 3). Globus may move
  into stage 4. AppLeS and NWS will move to stage
  2.
• NPACI, NCSA, IPG, DoD HPCMOD
• Tutorials yesterday, more to come.
• By 2000-2001, both Legion and Globus should be in
  stage 4 and 5, integrated with AppLeS and NWS.

12
(No Transcript)
13
Engineering

• Provide infrastructure to enable large-scale
  simulations of multi-scale, multi-physics
  phenomena.
• Use simulation tools to design new materials,
  predict effects of subsurface pollutants, assess
  productivity of oil reservoirs, and design large,
  electromagnetic communication systems.
• Integrate simulation tools, imaging technologies,
  and modern computational technologies across
  distributed compute platforms (DICE, IE, META).
• Integrate data-intensive computing technologies
  into simulation tools to allow tomography data to
  be incorporated into analysis of complex
  engineering systems (PTE, DICE, IE).
• Provide hardened tools to the scientific
  community.
• Develop parallel versions of important
  engineering applications (META, PTE, IE, RWG).
• Develop new algorithms that function in parallel
  adaptive environments to enhance performance of
  legacy codes (META, RWG).

14
Programming Tools Environments

• Building blocks to enable modeling of complex
  physical phenomena.
• Provide library solutions for parallel numeric
  and run-time issues.
• Parallel methods for sparse, nonsymmetric linear
  systems, finite-element problems, n-body
  problems, and eigenproblems.
• Improve/deploy KeLP run-time adaptive library
  use ESS problem as a testbed.
• MetaChaos library to support interoperability of
  separately developed parallel programs scenarios
  from multiple thrusts.
• Harden language and compiler technology.
• Develop compiler technology to restructure
  applications programs to improve large-scale I/O
  performance, initially for an ESS application.
• Active Data Repository to enable efficient
  querying/processing of very large, complex
  multi-dimensional data sets.
• Integration with SRB use in ESS, NS, and ENG
  testbeds.

15
Molecular Science

• Understand how fundamental molecular properties
  contribute to macroscopic phenomena in chemistry
  and biology.
• Simulate molecular dynamics for large systems
  (e.g., biological molecules).
• Port existing codes to parallel machines, test
  them, and apply to problems not currently within
  reach (CR, MS, PTE).
• Create databases for molecular systems to support
  exploratory analysis, hypothesis generation,
  communication, dissemination.
• Create and populate data schema for critical
  areas Biological macromolecules, MD
  trajectories, quantum computations (DICE, PTE).
• Create visualization technologies for
  communication/analysis (IE, MS).
• Provide hardened tools to scientific community
  for use.
• Identify critical algorithms requiring HPC,
  implement on NPACI hardware.
• Conduct education, outreach, and training of
  scientists/students (EOT, IE, CR).

16
Interaction Environments

• Provide a customizable collaborative scientific
  computing environment to communicate knowledge,
  ideas, and data to facilitate cross-disciplinary
  research.
• Enable broad-based collaborative environments by
  integrating collaboration control
  infrastructure, visualization tools, data access,
  remote instrument control, and execution-level
  environment (DICE, META, PTE).
• Enable creation and dissemination of
  visualizations and animation showing the latest
  scientific results to a broad community of
  students, teachers, and the public (EOT).

17
Interaction Environments (cont.)

• Broad-based Collaborative Environment
• Define and implement collaboration-control
  architecture to monitor collaboration services.
• Select, harden, and integrate a set of
  visualization and virtual-reality tools that
  provide broad capabilities to represent and
  communicate scientific data.
• Provide user-interface tools to access remote
  data archives and online real-time data
  acquisition from remote sensors support large
  data/ large simulation visualization.
• Provide high-quality, network-based video for
  remote control of scientific instruments (NS).
• Demonstrate/evaluate utility in Earth System
  Science prototype.
• Demonstrate/evaluate utility in Molecular Science
  prototype.

18
Education, Outreach, and Training

• Provide infrastructure to ensure all citizens may
  make productive use of emerging computing
  technologies to advance their ability to
  understand/solve problems in education, science,
  business, government, and society.
• Have a national impact on computational science
  education (K-12, undergrad, grad, informal).
• Extend outreach of PACI technologies to new user
  communities (social sciences, humanities).
• Provide access/inclusion in computational science
  to underrepresented communities (women,
  minorities, people with disabilities).
• Develop new learning technologies/deploy them on
  PACI infrastructure.
• Scale exemplary programs to a national level
  using formal evaluation.

19
Maturation of the Visions

• Maintain focus on enabling important scientific
  understanding.
• Create more linkages among thrusts.
• Plan demonstration projects within next year.
• Identify specific implementation steps.
• Identify timeframes for deliverables.

20
Where We Go from Here...

• We need your help to refine these visions!
• Contact relevant thrust area leaders (a list of
  leaders is in your registration packet and the
  NPACI Partnership Report).
• This will be an ONGOING process!
• Progress will be noted on the NPACI intranet.
• Visions and software roadmap strategy (next
  session) will be combined in a design document.
• This process will guide future activities and
  funding decisions!
• Process endorsed by NSF Program Review Panel.