Open nv-rliu opened 3 weeks ago
Occurs when using netscience without renumbering
Reproducing strategy: take the non-renumbered data from Python and convert it to C++ vectors to see what the algorithm is doing and why it is running into a memory error
Version
24.10
Which installation method(s) does this occur on?
Conda
Describe the bug.
When running
tests/test_k_truss_subgraph_mg.py
on 2-GPU ondraco-rno
, the test encounters a memory error that causes it to fail.Minimum reproducible example
Relevant log output
Environment details
Other/Misc.
Was unable to reproduce this failure on the lab machines. Also, this failure can be seen without running the entire suite of cugraph MG tests inside an interactive slurm session.
Code of Conduct