A Julia implementation of Linda for parallel programming?


#1

I wonder if anyone here has heard of Linda and has an opinion on whether Julia should implement it in the future? I’m not qualified to have an opinion; I’m just throwing the idea out there for those who are.

In this blog article, the author compares Linda to the metric system and MPI to doing physics with foot-acres. Linda certainly does seem simpler, so I guess the only question is how it would compare performance-wise. If there wouldn’t be a performance sacrifice, then it seems like Linda would be a great addition to Julia.

Thoughts?


#2

cf the now apparently defunct


#3

It looks attractive to me as it seems to avoid any distinction between multiple threads and multiple processes. Which means I could focus on the code, forget about the background implementation, and let Julia take care of the details.