Julia comparison

[jlapeyre]

The Julia graphs people have done benchmarks comparing in the past (I think with code) the package LightGraphs.jl (which was renamed Graphs.jl a couple of weeks ago) to several other libraries. LightGraphs.jl compared very favorably to other graph packages.

However, when I compared betweenness centrality between Julia and retworkx, I found retworkx was significantly faster, something like 50% or more IIRC.

It might be worth adding Julia Graphs.jl benchmarks here

[mtreinish]

Yeah, I'm good with adding comparisons to julia's graph library here. Right now the repo is scoped with just hosting the benchmarks we're using for our pending paper submission to JOSS which we're drafting here: Qiskit/rustworkx#455 which is mostly concerned about how retworkx compares to other python libraries.

But I think it's a good idea to add other graph libraries here and expand the benchmarks to other routines than just what we use in the paper.

[IvanIsCoding]

If you are willing to implement the benchmarks for Julia, it would be a great addition. I do not know if we will have space on the paper, but I think broadening the scope of this repo is good. It would let us check how retwork is comparing periodically.

The Julia graphs people have done benchmarks comparing in the past (I think with code) the package LightGraphs.jl (which was renamed Graphs.jl a couple of weeks ago) to several other libraries. LightGraphs.jl compared very favorably to other graph packages.

However, when I compared betweenness centrality between Julia and retworkx, I found retworkx was significantly faster, something like 50% or more IIRC.

It might be worth adding Julia Graphs.jl benchmarks here

Do you have the link to that benchmark you mentioned? Maybe we could ask for retworkx to be included there

[cyrusmsk]

I'm not sure which benchmark was supposed by topic starter, but there is one benchmark which also has Julia's solution - https://www.timlrx.com/blog/benchmark-of-popular-graph-network-packages-v2