optimization: allow a single open/create call for N parallel processes
Nobody has claimed this yet.
- Dominant language
- C
- Stars
- 122
- Forks
- 34
- PR merge metrics
- No merged PRs in 30d
Description
Extracting from a discussion on the mailing list:
> At the moment, the other clients (rank >= 1) still need to call unifyfs_open(). Doing so will initialize some internal client state for the target file that is required by the I/O dispatch.
> However, I think it would be reasonable for Unify to support relaxing this so that your example could work as is. We can easily do a metadata lookup using just the gfid provided in the I/O req.
In ROMIO, we have seen file systems that provide a stateless global handle allow for really scaleable metadata (I am always excited when I can cite our 17 year old paper) https://link.springer.com/chapter/10.1007/978-3-540-30218-6_18
Not critical: ROMIO of course can handle calling open from every process just fine.
Contributor guide
No contributing guide indexed for this repository
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
The issue names unifyfs_open(), the I/O dispatch, and metadata lookup by gfid, but no files or tests. Start by locating those entry points and determine the design and validation needed for rank >= 1 processes to issue I/O without calling open; done means the single open/create workflow works correctly for N parallel processes.
Written by the indexing model from the issue text.
Assessment
- Tech stack
- c
- Domain
- distributed-systems, performance
- Issue type
- Feature
- Difficulty
- 5/5
- Estimated time
- Over a week
- Activity status
- Stale
- Clarity
- Needs clarification
- Newbie friendliness
- 30/100