Community detection running times

[jeshi96]

Hi,

I'm testing out some of the example applications (specifically examples/community, examples/cliques, and examples/kcores), and I wanted to make sure that the running times that I'm getting are expected, especially for community detection.

I'm using a GCP machine (c2-standard-60) with 30 cores with 2-way hyper-threading (60 vCPUs) and 240 GB of memory. The two graphs that I'm running on are the SNAP datasets com-DBLP and com-Youtube.

In 4 hours, I am unable to finish community (for both -a:1 and -a:2, so both Girvan-Newman and Clauset-Newman-Moore) for both com-DBLP and com-Youtube.

For cliquesmain, using -k:2, com-DBLP completes in 1 hour 19 minutes, but com-Youtube does not finish in 4 hours.

For kcores, com-DBLP completes in 0.72 seconds and com-Youtube completes in 1.94 seconds.

Are these running times expected (particularly for community detection, given that it's unable to finish)? Thanks!

Best,
Jessica