[BUG] ANN_BENCH: the data/query memory types not checked in multi-gpu implementations
Nobody has claimed this yet.
- Dominant language
- Cuda
- Stars
- 854
- Forks
- 236
- Avg merge
- 3d 3h
- Merged PRs (30d)
- 62
Description
All multi-gpu index wrappers do not check the pointer types and wrap them into host mdarray views:
https://github.com/rapidsai/cuvs/blob/73e4fbcb981a67048a7f63b39416a87cc878c1ba/cpp/bench/ann/src/cuvs/cuvs_mg_ivf_pq_wrapper.h#L125-L138
The pointer locations are configurable in the benchmarks at runtime, hence we should check them and, at the very least, throw an error if the memory location is not supported.
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 with cuvs/cuvs_mg_ivf_pq_wrapper.h around lines 125-138, then inspect the other multi-GPU index wrappers for the same pointer handling. Check how benchmark memory locations are configured at runtime and verify that unsupported locations produce an error instead of being wrapped as host mdarray views.
Written by the indexing model from the issue text.
Assessment
- Tech stack
- cpp
- Domain
- performance
- Issue type
- Bug
- Difficulty
- 3/5
- Estimated time
- 1-2 days
- Activity status
- Stale
- Clarity
- Mostly clear
- Newbie friendliness
- 48/100