interesting read about high performance concurrency & networking
I just discovered http://www.seastar-project.org/ an interesting open source C++ project for high performance message passing based parallelism between cores & for high speed networking. It came out of http://osv.io/ (a particular example of the whole "unikernels" movement,. Erlang on Xen is another).
The programming model is similar to Erlang but all C++ based. Could be interesting just as inspiration or even as a native lib to base a future implementation of single process parallelism in Julia on.
I wonder how this would compare in performance to using MPI for multicore parallelism (on a single host).