search

puzzlef / leiden-communities-openmp

Design of OpenMP-based Parallel Leiden algorithm for community detection.
https://arxiv.org/abs/2312.13936
MIT License
11 stars 1 forks source link

Cannot see any results in log #3

Closed jgbradley1 closed 2 weeks ago

jgbradley1 commented 3 weeks ago

I am experiencing difficulty in running the main.sh script to reproduce the results of running this parallel version of Leiden (the results in the paper sound great though!).

With minor modifications to the main.sh script, I added a devcontainer to the repo in a fork here to have a more reproducible environment that I can run experiments with on a laptop (with fewer threads defined) for initial testing.

I downloaded the europe_osm.mtx graph data and when I execute main.sh, I see messages in the log file (see below) but it's not the output I expect. Based on this line of code, I was expecting to see more informative output that describes the discovered partitions. I see similar output when I run on one of the other graph datasets referenced in the paper.

LOG FILE

2024-11-05 21:50:26 OMP_NUM_THREADS=4
2024-11-05 21:50:26 Loading graph /root/Data/europe_osm.mtx ...
2024-11-05 21:50:41 order: 50912018 size: 108109320 [directed] {}
2024-11-05 21:51:35 OMP_NUM_THREADS=4
2024-11-05 21:51:35 Loading graph /root/Data/europe_osm.mtx ...
2024-11-05 21:51:45 order: 50912018 size: 108109320 [directed] {}
2024-11-05 21:52:30 OMP_NUM_THREADS=4
2024-11-05 21:52:30 Loading graph /root/Data/europe_osm.mtx ...
2024-11-05 21:52:44 order: 50912018 size: 108109320 [directed] {}
2024-11-05 21:53:26 OMP_NUM_THREADS=4
2024-11-05 21:53:26 Loading graph /root/Data/europe_osm.mtx ...
2024-11-05 21:53:39 order: 50912018 size: 108109320 [directed] {}
2024-11-05 21:54:31 OMP_NUM_THREADS=4
2024-11-05 21:54:31 Loading graph /root/Data/europe_osm.mtx ...
2024-11-05 21:54:45 order: 50912018 size: 108109320 [directed] {}

Note: I am trying to run the code from the main branch. It's not clear why the code appears to fail silently or if it's running as expected. Can you provide any suggestions on what could be going wrong?

jgbradley1 commented 2 weeks ago

Will close this issue for now. I discovered the program was getting killed due to OOM issues.