Seunghwa Kang
Seunghwa Kang
I am seeing a similar issue with `thrust::remove` & `thrust::remove_if`. Sounds related and better be fixed together.
The new triangle counting works only on undirected graphs, and it will throw an exception instead of returning an inaccurate result. https://github.com/rapidsai/cugraph/blob/branch-22.06/cpp/src/community/triangle_count_impl.cuh#L142
And I am not sure NetworkX really supports triangle counting on directed graphs. See ``` @not_implemented_for("directed") ``` https://github.com/networkx/networkx/blob/58b63cb57cd1747c23611ee0b46991a5be2db751/networkx/algorithms/cluster.py#L19 And ``` >>> import networkx as nx >>> G = nx.DiGraph() >>>...
To support triangle counting on directed graphs, we should define what triangles mean in directed graphs first. Should we consider only a->b, b->c, c->a or is a->b, a->c, b->c also...
libcugraph part is complete (https://github.com/rapidsai/cugraph/blob/branch-22.04/cpp/include/cugraph/algorithms.hpp#L1533).
With this proposed abstraction, can we use this new feature in .cpp files (instead of requiring nvcc); we're forced to use rmm::device_buffer instead of rmm::device_vector in several places as the...
> > With this proposed abstraction, can we use this new feature in .cpp files (instead of requiring nvcc); we're forced to use rmm::device_buffer instead of rmm::device_vector in several places...
> ``` > fill(data(), size, fill_value); // this function fills [data(), data() + size * fill_value.size()] by repeating fill_value size times > ``` > > I'm confused by your example....