Eugenio Angriman

@nahuelalmeira the [NetworKit Development Guide](https://networkit.github.io/dev-docs/DevGuide.html#devGuide) provides all the information you need to know about how to develop and send contributions to NetworKit. Then a possible starting point would be to...

I don't have detailed benchmark data. I can imagine that, on smaller graphs than the ones I am considering, the dynamic schedule has more synchronization costs so it performs slightly...

We could have a more generic function that takes a distribution: ```cpp template void GraphTools:generateRandomWeights(Graph &G, Distribution distr = std::uniform_real_distribution{0, 1}) { if (!G.isWeighted()) G = GraphTools::toWeighted(G); G.parallelForEdges([&](node u, node...

@michitux @larsgottesbueren are there any updates on this PR? Is it possible to split it into smaller contributions?

Hi, thank you for the PR! About your questions: 1. Yes that is fine 2. Same, we do that as well in other parts of the codebase 3. I guess...

I replicated your issue and, after some testing, it seems that `ParallelConnectedComponents` is bugged. Our tests are not failing because`getComponents()` is never tested for this algorithm. My recommendation is to...

I don't think we should generate a multigraph, they are not well-supported in NetworKit. What about taking as input a binary operator? We could use `std::plus` by default. Edit: this...

This seems quite simple to do; we could use an internal variable that suppresses the warning.

The segfault occurs because the NetworKit binary writer does not support empty graphs. We'll try to fix this asap.

Here is an example of performance benchmarks of popular graph/network analysis frameworks including NetworKit and networkX: https://www.timlrx.com/blog/benchmark-of-popular-graph-network-packages