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!
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