[QST] Graph Construction time using CAGRA

Open
#1,963 2 comments 0 reactions 0 assignees View on GitHub

Nobody has claimed this yet.

Assessment

Difficulty
4/5
Estimated time
3-5 days
Newbie friendliness
25/100
Issue type
Bug
Clarity
Needs clarification
Activity status
Stale
Tech stack
python
Domain
performance, search

Research direction

Review the CAGRA Python build API usage shown in the issue screenshots and compare the reported SIFT1M RTX3090 timings with the CAGRA paper. Check how nndescent and RMM resource allocation are configured; done means determining whether the timing is expected or identifying a reproducible API misuse.

Written by the indexing model from the issue text.

Description

question

Hi,

I'm using cagra build python api to perform ann search.

However, I found that build time is extremely slow even if I use nndecent and rmm resource allocation. Below is my code
image
image

I tested sift1m dataset on RTX3090 GPU, does the performance make sense? Because in cagra paper, it says building time is fast.
image

I'm not sure if I have make some mistakes in using this api?

Dominant language
Cuda
Stars
1k
Forks
251
Avg merge
1d 8h
Merged PRs (30d)
31

Contributor guide

No contributing guide indexed for this repository

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.

More from NVIDIA/raft

All issues in NVIDIA/raft

Similar issues

More Performance issues

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.