XcalableMP Handbook

This website introduces a parallel language XcalableMP.

The explanation is based on the latest specification, but some features from the upcoming one are also included.

• 1. Introduction
  • 1.1. Background
  • 1.2. Features
• 2. Tutorial (Global-view)
  • 2.1. Introduction
  • 2.2. Data mapping
  • 2.3. Work mapping
  • 2.4. Inter-node communication
• 3. Tutorial (Local-view)
  • 3.1. Introduction
  • 3.2. Coarray declaration
  • 3.3. One-sided communication
  • 3.4. Synchronization
  • 3.5. Tutorial
• 4. nodes directive
  • 4.1. 1-dimensional node set
  • 4.2. Multi-dimensional node set
  • 4.3. Dynamic node set
  • 4.4. Partial node set
• 5. template directive
  • 5.1. 1-dimensional template
  • 5.2. Multi-dimensional template
  • 5.3. Dynamic template
• 6. distribute directive
  • 6.1. block distribution
  • 6.2. cyclic distribution
  • 6.3. block-cyclic distribution
  • 6.4. gblock distribution
  • 6.5. Distribution of multi-dimensional template
• 7. align directive
  • 7.1. Normal alignment
  • 7.2. Special alignment
• 8. Dynamic allocation of distributed array
  • 8.1. One-dimensional array
  • 8.2. Multi-dimensional array
• 9. template_fix construct
• 10. task/tasks construct
  • 10.1. task construct
  • 10.2. tasks construct
• 11. loop construct
  • 11.1. Accessing distributed array
  • 11.2. Reduction computation
  • 11.3. Parallelizing nested loop
• 12. array construct
• 13. shadow directive / reflect construct
  • 13.1. Declare shadow
  • 13.2. Update shadow
  • 13.3. Multi-dimensional shadow
• 14. gmove construct
  • 14.1. Collective mode
  • 14.2. in mode
  • 14.3. out mode
• 15. barrier construct
• 16. reduction construct
• 17. bcast construct
• 18. wait_async construct
• 19. reduce_shadow construct
• 20. coarray notation
  • 20.1. Synchronization statements
  • 20.2. Arguments
• 21. post/wait construct
• 22. lock/unlock construct
• 23. Procedure interface
• 24. Cooperation with OpenMP
• 25. Cooperation with MPI
  • 25.1. Calling MPI program from XMP program
  • 25.2. Calling XMP program from MPI program
• 26. Cooperation with Python
  • 26.1. Calling from parallel Python program
  • 26.2. Calling from sequential Python program