Wiki » History » Version 3

« Previous - Version 3/39 (diff) - Next » - Current version
Brian Nord, 01/11/2013 05:26 PM

Simulation Pipeline

Working Group Organization

  • Organizers: Nord, Gamper, Hambrecht, Amara, Refregier
  • Wrapper: Hambrecht, Nord, Farrens
  • Instrument Design: Kron
  • Target Catalog: Jouvel, Abdalla
  • Survey Strategy: Serrano, Nord, Annis
  • Fiber Allocation: Forero, Nord
  • Throughput: Nord, Saunders
  • Signal-to-noise: Coles, Cunha, Abdalla
  • Redshift Estimation: Coles, Cunha, Abdalla
  • Science: Kirk, Amara, Gaztanaga, Nichol, Beynon


Subsystem Organization:

  • Passing inputs/outputs:
    • Plan for organizing the pipeline:
      The following elements are listed next to each variable in "Inputs/Outputs" section of each Subsystem
    • Each subsystem has inputs and outputs: galaxy spectra and counts are the variables that are fed through each subsystem. In the examples below (and on each subsystem page):, the first input in the list is the variable (spectrum) and others are parameters (which might be wholly independent from other subsystems or come from another).
    • Each subsystem has input parameters with priors
    • Each subsystem receives from another and sends to another
    • Examples:
      • Input: item [ from where ] [prior]
      • Output: item [ to where ]
  • Technical details of the pipeline: specifics for workflow components
    • parameter and data file types
    • codes to be used
    • workflow design
      • References: attachment:Zuntz-PE-framework.pdf and attachment:zuntz-pe-framework-details.pdf
    • Where will we run the pipeline
      • at FNAL? ... code repository already near to existing DES-related infrastructure; no additional logins required
      • at ETH? ... available effort and expertise

Codes and Languages

Subsystem Language New Language Maintainer
Target Acquisition Supermongo C? S. Jouvel
Tiling IDL Python Nord
Survey Strategy Tcl Python Nord
Throughput Python n/a Nord
Signal-to-noise C n/a Jouvel
Redshift C n/a Oliver
DE FoM Matlab n/a Kirk



Preliminary set of milestones

  • There are pre-Snowmass and Snowmass meetings in March(ish) and July(ish), so we want to have this pipeline many times by May/June at the latest.
  • At December Texas Meeting
    1. get last i/o params for each module
    2. Establish timeline for sets of runs
    3. establish priorities of what sim pipeline should say about instrument design
    4. Decide which trade studies we will do -- what kind of parameter variations will we perform?-
    5. Think about pre-DESpec deployment projects/publications
    6. Sign up for code development
    7. Clarify exact flow of pipeline
    8. Decide on which languages to use.
    9. each group should identify some quality assessment tests