Moves more MG graph ETL to libcugraph and re-enables MG tests in CI #3941
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
rapids-bot bot
pushed a commit
that referenced
this pull request
Oct 17, 2023
This PR temporarily disables single-GPU "MG" tests in CI until the dask related transient failures are resolved. The PR to re-enable them, intended to block the 23.12 release, is [here](#3941). Authors: - Joseph Nke (https://github.com/jnke2016) Approvers: - Rick Ratzel (https://github.com/rlratzel) - Ray Douglass (https://github.com/raydouglass) URL: #3940
divyegala
pushed a commit
to divyegala/cugraph
that referenced
this pull request
Oct 18, 2023
This PR temporarily disables single-GPU "MG" tests in CI until the dask related transient failures are resolved. The PR to re-enable them, intended to block the 23.12 release, is [here](rapidsai#3941). Authors: - Joseph Nke (https://github.com/jnke2016) Approvers: - Rick Ratzel (https://github.com/rlratzel) - Ray Douglass (https://github.com/raydouglass) URL: rapidsai#3940
… for graph creation (SG and MG)
…memory footprint when graphs are large relative to GPU memory
…to branch-23.12_re-enable-mg-testing
|
/ok to test |
|
/ok to test |
rlratzel
changed the title
|
Added a |
AyodeAwe
approved these changes
Nov 22, 2023
ChuckHastings
approved these changes
Nov 22, 2023
|
/ok to test |
|
/ok to test |
rlratzel
reviewed
Nov 26, 2023
| weight_array=None, | ||
| store_transposed=False, | ||
| renumber=False, | ||
| do_expensive_check=False, | ||
| edge_id_array=None, | ||
| edge_type_array=None, | ||
| input_array_format="COO", | ||
| vertices_array=None, |
There was a problem hiding this comment.
This (and also the updated MGGraph ctor) still does not have all new args at the end, but maybe since only drop_self_loops and drop_multi_edges are affected, it's not that likely to break users using only positional params? I'll approve it but I'll also label this PR as a breaking change.
rlratzel
approved these changes
Nov 26, 2023
rlratzel
added
breaking
|
/ok to test |
|
/merge |
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.
This PR includes changes that moves some of the MG graph etl steps (such as computing number of edges) to libcugraph to reduce the amount of dask overhead involved in graph creation. Those ETL steps were also responsible for various dask-related transient errors that caused us to temporarily disable MG testing in CI. These changes allow us to re-enable MG testing in CI, so this PR includes that update too.
closes #3948