Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

[FEA][NV] Move nvgraph repo in cugraph repo #33

Closed
8 of 9 tasks
afender opened this issue Jan 4, 2019 · 3 comments
Closed
8 of 9 tasks

[FEA][NV] Move nvgraph repo in cugraph repo #33

afender opened this issue Jan 4, 2019 · 3 comments
Assignees
Labels
feature request New feature or request

Comments

@afender
Copy link
Member

afender commented Jan 4, 2019

This task will result in open sourcing nvGraph. NvGraph contains the core CUDA code for Jaccard, Louvain, BFS, SSSP and more.

The goal is to open source all at once to facilitate the integration and minimize the refactoring effort.

Before moving nvGraph repo here, the following tasks must be done in nvGraph:

Then :

  • Copy nvgraph code in cugraph repo: code should be in a nvgraph folder under cpp/nvgraph
  • Add build option to produce nvgraph library
  • rename nvgraph stand alone library nvgraph_st into nvgraph_rapids
  • Update cugraph infrastructure to build the newly added nvpraph code

Important improvement

  • Remove multiple CUB dependencies (3 versions of cub in cugraph after transferring nvgraph...)
  • Reduce build time (nvgraph takes minutes to build, this will negatively impact cugraph development)
@afender afender added feature request New feature or request 0 - Backlog labels Jan 4, 2019
@afender afender changed the title Move nvgraph repo in cugraph repo [FEA][NV] Move nvgraph repo in cugraph repo Jan 4, 2019
@BradReesWork BradReesWork added this to the 0.11.0 milestone Jan 8, 2019
dantegd pushed a commit to dantegd/cugraph that referenced this issue Jan 18, 2019
[REVIEW] Added ability to call cuML with numpy arrays and general test cleaning
@afender
Copy link
Member Author

afender commented Feb 7, 2019

Open sourcing now would reduce the deployment overhead and remove an unnecessary complexity layer in cugraph. Currently, nvgraph has to be updated on anaconda.com each time something change which is a source of bugs and slows down development.

@BradReesWork BradReesWork removed this from the 0.11.0 milestone Feb 28, 2019
@afender
Copy link
Member Author

afender commented Mar 29, 2019

I was able to move nvgraph in cugraph repo and successfully completed the subtasks above. Nvgraph is automatically built with cugrph but can also be built as a stand-alone. It still takes at least 1 minute to build after optimizations. This works locally but Jenkins returns errors inside cub currently

@afender
Copy link
Member Author

afender commented Apr 8, 2019

closed by #194

@afender afender closed this as completed Apr 8, 2019
BradReesWork pushed a commit that referenced this issue Jan 23, 2020
BradReesWork pushed a commit that referenced this issue Nov 23, 2020
Forked branch-0.16 updated from Release branch-0.16
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
feature request New feature or request
Projects
None yet
Development

No branches or pull requests

3 participants