NVIDIA / NVIDIA/cuvs

[FEA] Support int64_t nearest neighbor index type for all ANN algorithm.

Open
#156 1 comment 0 reactions 0 assignees View on GitHub

Nobody has claimed this yet.

feature request
Dominant language
Cuda
Stars
854
Forks
236
Avg merge
3d 3h
Merged PRs (30d)
62

Description

Currently most of our the nearest neighbors methods support "int64_t" type for the nearest neighbor indices. The only exception is CAGRA, that returns uint32_t indices. For consistency, it would be great to support int64_t for CAGRA as well, and it can be done with a small overhead by adding a simple mapping after the search.

Similarly, it would be easy to add uint32_t support for all the other nearest neighbor algorithms, if there is a need for that.

Contributor guide

Open the contributing guide

First steps

  1. Read the whole issue, then the project's contributing guide.
  2. Comment on the issue to say you are picking it up — it saves two people doing the same work.
  3. Fork the repository and make your change on a branch.
  4. Open a pull request that references the issue number.

Research direction

Start by tracing the CAGRA nearest-neighbor search path and compare how other ANN algorithms represent int64_t indices. The work is done when CAGRA provides int64_t indices consistently with the other methods, with the proposed mapping overhead validated for correctness.

Written by the indexing model from the issue text.

Assessment

Domain
search
Issue type
Feature
Difficulty
4/5
Estimated time
3-5 days
Activity status
Stale
Clarity
Mostly clear
Newbie friendliness
35/100

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.