Hello All,
I'm using Aura Data Science to run graph algorithms such as Betweenness and Closeness Centrality on a graph with 5 million nodes and 85 million edges. Aura recommended a configuration with 2 CPUs, 8 GB of memory, and 16 GB of storage, costing $2 per hour. However, when I execute these algorithms, they run for hours without producing any results, despite working well for PageRank and Eigenvector Centrality. I even scaled up to the maximum of 20 CPUs, but the algorithms still run indefinitely and eventually terminate without any outcomes. What might I be doing wrong?
Additionally, if these algorithms are complex to run, why did Aura suggest a configuration with only 2 CPUs?