2015 06 03

• Agenda:
  • open TODOs:
    • DONE MT: create ticket for long-run problem
    • TODO MT: move semantic elements of tools into RP.utils
    • TODO MS: add/move remaining pieces (dataframe for profiles)
    • TODO AM: proposal to json export to persistent storage
    • TODO MS: proposal for persistent experimental data storage
    • DONE AM: create exp repo name space (github and bitbucket)
    • DONE AM: document different github / bitbucket projects
  • Development Progress:
    • None (paper writing)
  • Data Roadmap:
    • document
    • TODO RP: start iterating on the API
    • TODO MS: look over the spark scripts etc.
  • Experiments:
    • canonical problems for experiments
      • /bin/true, /bin/sleep, /bin/dd obviously
      • Gromacs
        • we can tune iteration numbers and number of cores
        • how do we tune data/memory sizes
        • mem is defined by physical model, which is defined in input files
        • TODO AM: create small problem examples in repo
        • TODO VB: add medium / large example
        • TODO AT: add Amber examples eventually
        • TODO MS: add NAMD examples eventually? (Tom Bishop)
        • Gromacs is ok up to one node granularity
      • Synapse:
        • we can tune EVERYTHING! :)
        • TODO AM: set up RP example
    • MT/MS: experimental data management service + tool to export data to that service after run
      • radical-update <session-id> -name my_experiment_123
      • 1. export json / collect profile / annotate metadata (TODO AM: proposal?)
      • 2. storage / provenance on some database / disk / service (TODO MS: proposal?)
      • 3. some interface to query and access
  • Publications:
    • RP Paper:
      • TODO ALL: commit plots and graphs etc., discuss on next call.
      • TODO AM: micro benchmark data plots
• Notes:
  • TODO AM: check documentation of state diagram in released docs