Commit e852b3c
authored
Hetrograph and dask_cudf support (#2592)
This PR fixes rapidsai/graph_dl#27
This PR fixes rapidsai/graph_dl#43
This PR fixes rapidsai/graph_dl#39
**Tests Added:**
Single GPU:
- [x] APIs like num_nodes, num_edges
- [x] test_sampling_basic
- [x] test_sampling_homogeneous_gs_in_dir
- [x] test_sampling_homogeneous_gs_out_dir
- [x] test_sampling_gs_homogeneous_neg_one_fanout
- [x] test_sampling_gs_heterogeneous_in_dir
- [x] test_sampling_gs_heterogeneous_out_dir
- [x] test_sampling_gs_heterogeneous_neg_one_fanout
Multi GPU:
- [x] APIs like num_nodes, num_edges
- [x] test_sampling_basic
- [x] test_sampling_homogeneous_gs_in_dir
- [x] test_sampling_homogeneous_gs_out_dir
- [x] test_sampling_gs_homogeneous_neg_one_fanout
- [x] test_sampling_gs_heterogeneous_in_dir
- [x] test_sampling_gs_heterogeneous_out_dir
- [x] test_sampling_gs_heterogeneous_neg_one_fanout
Bugs to reproduce:
- [ ] Repro heterogeneous single gpu hang outside pytest
- [x] Repro hetrogenous multi gpu incorrect results for with_replace=False
#2760
- [x] Repro hetrogenous incorrect results for different amount of GPUs
#2761
Tests that depend upon #2523
- [x] Add minimal example to PR to ensure it gets fixed
Added comment here: #2523 (review)
- [x] test_get_node_storage_gs (Failing cause of a PG bug)
- [x] test_get_edge_storage_gs (Failing cause of a PG bug)
- [x] test_get_node_storage_gs (Failing cause of a PG bug)
- [x] test_get_edge_storage_gs (Failing cause of a PG bug)
Authors:
- Vibhu Jawa (https://github.com/VibhuJawa)
Approvers:
- Rick Ratzel (https://github.com/rlratzel)
- Brad Rees (https://github.com/BradReesWork)
URL: #25921 parent a299645 commit e852b3c
File tree
3 files changed
+1108
-342
lines changed- python/cugraph/cugraph
- gnn
- tests
- mg
3 files changed
+1108
-342
lines changed
0 commit comments