PMIx Reference Run Time Environment
  • 1. Quick start
  • 2. Release Notes
  • 3. Getting help
  • 4. Installing PRRTE
  • 5. PRRTE DVM Configuration
  • 6. Terminology
  • 7. How Things Work
  • 8. Implementation Plans
    • 8.1. Mapping by App Plans
    • 8.2. Mapping by Device Plans
    • 8.3. Node Reservation and Session Targeting Plans
    • 8.4. Elastic DVM Plans
    • 8.5. DVM Bootstrap Plans
    • 8.6. Group Fault-Tolerance Plans
    • 8.7. Scalable Collectives Plans
    • 8.8. Cross-DVM Data Server Plans
  • 9. Host specification
  • 10. Process placement
  • 11. Launching applications
  • 12. Notifications
  • 13. Session directory
  • 14. Developer’s guide
  • 15. Contributing to PRRTE
  • 16. Deferred work
  • 17. PRRTE Community Code of Conduct
  • 18. License
  • 19. PRRTE manual pages
  • 20. Software Version Numbers
  • 21. Version Compatibility Within a DVM
  • 22. News
PMIx Reference Run Time Environment
  • 8. Implementation Plans
  • View page source

8. Implementation Plans

Detailed implementation plans for in-progress and upcoming features.

  • 8.1. Mapping by App Plans
    • 8.1.1. Per-App-Context Mapping Policies
    • 8.1.2. Per-App-Context Mapping: Implementation Plan
    • 8.1.3. Per-App-Context Mapping: Implementation Tasks
  • 8.2. Mapping by Device Plans
    • 8.2.1. Mapping and binding by device
    • 8.2.2. Device mapping: implementation plan
  • 8.3. Node Reservation and Session Targeting Plans
    • 8.3.1. Node Reservation and Session Targeting: Specification
    • 8.3.2. Node Reservation and Session Targeting
  • 8.4. Elastic DVM Plans
    • 8.4.1. Elastic DVM: Specification
    • 8.4.2. Elastic DVM Implementation Plan
    • 8.4.3. DVM Grow-Campaign Fence Tracking
    • 8.4.4. DVM Shrink-Campaign Fence Tracking
    • 8.4.5. Collective Shrink Completion — Repair the Routing Tree Once per Campaign
  • 8.5. DVM Bootstrap Plans
    • 8.5.1. DVM Bootstrap: Specification
    • 8.5.2. DVM Bootstrap Implementation Plan
    • 8.5.3. Rewiring a Returned Daemon: the “Unheal” Path
  • 8.6. Group Fault-Tolerance Plans
    • 8.6.1. Degraded-mode survival of a group construct on daemon loss
  • 8.7. Scalable Collectives Plans
    • 8.7.1. Restructuring the DVM collectives around operations
  • 8.8. Cross-DVM Data Server Plans
    • 8.8.1. Reaching a data server in another DVM
Previous Next

© Copyright 2003-2026, The PRRTE Community. Last updated on 2026-09-04 02:47:49 UTC.

Built with Sphinx using a theme provided by Read the Docs.