References in JOSS paper

[landreman]

For the JOSS review:

Googling around quickly, it looks like another project to connect mpi with jax is https://github.com/mpi4jax/mpi4jax. How about citing their paper and adding a few words in your manuscript about the similarities/differences?

Also some info is missing on the Scott reference.

[AlanPearl]

• The Scott reference is now fixed: f9e0146
• Thanks for bringing up mpi4jax, I will definitely cite it and add some clarification on why to use one or the other...

[ewu63]

Chiming in on the JOSS review as well:

• I'm not familiar with jax in the context of distributed memory parallelism, but OpenMDAO from NASA Glenn is another computational framework that can 1) assemble derivatives of multiple components via the chain rule, and 2) work with parallel components under MPI. Might be worth referencing. I know you're basing the derivatives on jax which has some advantages which you can point out, such as efficient gradient computation via AD.
• Parallel PSO is an idea that has been implemented before. I don't know any references off the top of my head, so this is certainly not something you need to do, but any references to existing implementations I think would be helpful.

[AlanPearl]

I have pushed changes to the statement of need in a2b6e83 including references to similar past efforts, including the two mentioned in this thread. How does this look?

[landreman]

Looks good to me

[ewu63]

Good with me as well

[jbytecode]

Can we please close the resolved issues?

[AlanPearl]

Yes, this is now closed