MPI: Dynamically handle virtual node numbers (a.k.a work portions)
Nobody has claimed this yet.
- Dominant language
- C
- Stars
- 13.6k
- Forks
- 2.6k
- PR merge metrics
- No merged PRs in 30d
Description
@solardiz brought this up
I actually had a different thought regarding MPI: we might want to add an alternative MPI support mode for heterogeneous clusters where the number of virtual nodes would be much greater than the number of logical CPUs in the cluster, and processes would dynamically (via MPI messages) request/obtain next virtual node numbers after they've done their work with the previous number. (Like what I had posted to john-users with many masks and file renames, but instead with many nodes and MPI.)
Contributor guide
First steps
- Read the whole issue, then the project's contributing guide.
- Comment on the issue to say you are picking it up — it saves two people doing the same work.
- Fork the repository and make your change on a branch.
- Open a pull request that references the issue number.
Research direction
Start by reading the repository's existing MPI support and the linked pull request discussion, then trace how virtual node numbers and work portions are currently assigned. The issue provides no file or test entry point; done would require a defined design and working dynamic allocation through MPI messages for heterogeneous clusters.
Written by the indexing model from the issue text.
Assessment
- Tech stack
- c
- Domain
- distributed-systems
- Issue type
- Feature
- Difficulty
- 5/5
- Estimated time
- Over a week
- Activity status
- Stale
- Clarity
- Needs clarification
- Newbie friendliness
- 25/100