-
Notifications
You must be signed in to change notification settings - Fork 293
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
Multi gpu sample edges utilities #2064
Conversation
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
LGTM
Codecov Report
@@ Coverage Diff @@
## branch-22.04 #2064 +/- ##
===============================================
Coverage ? 73.55%
===============================================
Files ? 156
Lines ? 10295
Branches ? 0
===============================================
Hits ? 7573
Misses ? 2722
Partials ? 0 Continue to review full report at Codecov.
|
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
More review comments are coming.
rmm::device_uvector<gpu_t>& active_source_gpu_ids, | ||
EdgeIndexIterator edge_index_first, | ||
typename GraphViewType::vertex_type invalid_vertex_id, | ||
int indices_per_source, |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Better be size_t instead of int
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
I have changed it to edge type. We do not expect this value to exceed the maximum vertex degree.
adjacency_list_offsets.reserve(graph_view.get_number_of_local_adj_matrix_partitions()); | ||
adjacency_list_indices.reserve(graph_view.get_number_of_local_adj_matrix_partitions()); | ||
|
||
id_segments.push_back(0); |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
You can compute id_segments using get_vertex_partition_lasts() (https://github.com/rapidsai/cugraph/blob/branch-22.04/cpp/include/cugraph/graph_view.hpp#L115).
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
More reviews coming.
Moved files and renamed functions.
@gpucibot merge |
This PR tracks work on MNMG Neighborhood Sampling, for G(C)NN needs. Dependencies: 1. #1982 2. #2064 3. Integration of rapidsai/cugraph-ops#24 into `cugraph` Authors: - Andrei Schaffer (https://github.com/aschaffer) Approvers: - Seunghwa Kang (https://github.com/seunghwak) - Kumar Aatish (https://github.com/kaatish) - Chuck Hastings (https://github.com/ChuckHastings) - Rick Ratzel (https://github.com/rlratzel) URL: #2073
Add utilities to enable multi gpu gathering to be used for mnmg sampling.