[FEA] Enable using cugraph uniform sampling in multi client environments #3184
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
VibhuJawa
added
improvement
Codecov ReportBase: 55.22% // Head: 55.30% // Increases project coverage by
Additional details and impacted files@@ Coverage Diff @@
## branch-23.02 #3184 +/- ##
================================================
+ Coverage 55.22% 55.30% +0.07%
================================================
Files 148 148
Lines 9543 9573 +30
================================================
+ Hits 5270 5294 +24
- Misses 4273 4279 +6
Help us with your feedback. Take ten seconds to tell us how you rate us. Have a feature suggestion? Share it here. ☔ View full report at Codecov. |
|
That bug is also fixed in 3148 |
|
Could you add some documentation? Is setting |
|
Just a code comment is probably sufficient, maybe link to dask documentation. |
Done, added both. |
jnke2016
approved these changes
Jan 27, 2023
rlratzel
reviewed
Jan 27, 2023
python/cugraph/cugraph/dask/sampling/uniform_neighbor_sample.py
Outdated
Show resolved
Hide resolved
python/cugraph/cugraph/dask/sampling/uniform_neighbor_sample.py
Outdated
Show resolved
Hide resolved
python/cugraph/cugraph/dask/sampling/uniform_neighbor_sample.py
Outdated
Show resolved
Hide resolved
rlratzel
requested changes
Jan 27, 2023
|
/merge |
rapids-bot bot
pushed a commit
that referenced
this pull request
Jan 30, 2023
This PR makes `cugraph.structure.graph_classes` dask serializable and de-serializable. This means that we can cleanly use dask to transmit graph objects b/w different clients without much work and using all ready supported API. Depends on PR: ~#3184 (PR is IN) and closes #3193 See docs here: https://distributed.dask.org/en/stable/publish.html **On client 0** ```python3 print(type(g)) client.publish_dataset(mg_graph_for_sampling=g) ``` ``` <class 'cugraph.structure.graph_classes.MultiGraph'> ``` **On client 1** ```python3 g = client_worker_1.get_dataset('mg_graph_for_sampling') type(g) ``` ``` cugraph.structure.graph_classes.MultiGraph ``` CC: @rlratzel Authors: - Vibhu Jawa (https://github.com/VibhuJawa) Approvers: - Rick Ratzel (https://github.com/rlratzel) - Erik Welch (https://github.com/eriknw) URL: #3192
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
In multi client environments our sampling algorithms tend to hang, this PR adds a lock to prevent that from happening. This PR closes #3186
See notebooks here with MRE to reproduce.
Before PR (HANG):
After PR (NO HANG):
I have not added tests for this here because mocking dask with multi client environment is not possible in CI yet so we will do this as future work.
Also includes a bug fix in our sampling code path.