NVIDIA / NVIDIA/cuvs

Float codebook vectors support for PQ dataset and Iterative CAGRA-Q

Open
#2,590 0 comments 0 reactions 0 assignees View on GitHub

Nobody has claimed this yet.

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

Description

We currently support fp16 (half) codebook vectors types for PQ dataset and Iterative CAGRA-Q build.
The CAGRA-Q search kernels already support float codebook vectors types as well. We can explore whether it is worth implementing the float support end-to-end and if that has any benefits.

https://github.com/NVIDIA/cuvs/pull/1810#discussion_r3975137010

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 with the linked PR discussion, then trace the existing fp16 codebook-vector support for the PQ dataset and Iterative CAGRA-Q build, along with the CAGRA-Q search kernels' float support. Determine whether end-to-end float support provides benefits; done means reaching and documenting that decision and implementing the agreed scope if warranted.

Written by the indexing model from the issue text.

Assessment

Domain
search
Issue type
Feature
Difficulty
5/5
Estimated time
Over a week
Activity status
Active
Clarity
Needs clarification
Newbie friendliness
35/100

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.