B9: Astrophysics

1
B9 Astrophysics

• What can be achieved in the next 5 - 10 years ?
• Dark energy/dark matter data analysis (10
  experiments producing 10TB to 10PB datasets) -
  better understanding of the nature and evolution
  of the universe
• Multi-physics multi-scale simulations, e.g.
  supernovae, stellar-evolution, galaxy-formation
  and interaction, etc. - often spanning all
  physical interactions/forces (strong, weak, GR,
  EM, chemistry,etc) and gt 10 orders of magnitude
  in space/time.

2
B9 Astrophysics

• What are the challenges ?
• Management of data deluge (detectors Moore's
  law)
• Data access (technical sociological issues)
• Connecting observation simulation data
• Numerical precision
• Auto- cross- validation verification
• Algorithms implementations
• Quantifying statistical and systematic
  (including computational) uncertainties
• Interdisciplinary, interagency, international
  field
• Education training - this generation next

3
B9 Astrophysics

• What is the state of the art ?
• (Mostly) piecemeal by sub-field
• Code proliferation, much of it in the public
  domain
• Disparate data formats, models, types,
  distribution, provenance, accessibility
• How can we accelerate development ?
• "Decadal survey" for computational astrophysics
• Interagency coordination (including OMB),
  international outreach

4
B9 Astrophysics

• What are the expected outcomes and impacts of
  acceleration ?
• Simulation data standards cf. Virtual
  Observatory
• Concentrated effort on the most important
  computational astrophysical topics
• What scale of investment is needed ?
• Comparable to the experiments DoE is invested in
  already
• .SNAP/JDEM, LSST, DES, SDSS, etc.

5
B9 Astrophysics

• What are the risks ?
• Missed opportunities for optimizing experiments
• Building something no-one wants. This has to be
  science driven - no one machine/paradigm fits all
• Who/what are we missing ?
• Input from many from the field - many are at the
  APS meeting