2020 10 20.devel

• gpu sharing
• release planning
• configs, radical-base directory
  • use radical-base for config home
  • default_work_dir -> agent_radical_base
• cleaning of configs
• several pilots in one allocation
  • pilot partition
• MongoDB removal
  • feature/nodb
  • conda freeze?
• slate service @ ORNL
  • some issues lately
  • port numbers change for each deployment, and service dies now and then, so it's difficult to use
  • MT: no elastic IP/Port on slate?
    TODO HL: check with Shawn
    • why instability?
    • how to get stable port?
• MPI over jsrun
  • multiple GPUs not supported
  • TODO AM: ENTK #479
• problems with resource sharing between MPI ranks over jsrun?
  • TODO HL: provide reproducer
• RE: SMT
  • next week
• async workflow in MIDAS
  • example: double loop around bag of tasks
  • TODO AGENDA

TODOs

• resources
  • list all valid resources in docs?
    • TODO IP: check script
    • TODO AM: check resource configs (tacc / xsede)
• TODO AM: move radical-base to sandbox
• TODO AM: docs/architecture snippets: cleanup, integrate some
• TODO ALL: test coverage
• TODO ALL: populate project issues
  • AGENDA: discuss second Tuesday of month
• TODO AM: GTOD
• TODO AGENDA: switch from VE to CONDA?