Where's the nonblocking binding for MPI I/O?

[carlocastoldi]

Hi!
I hope this is not a question too dumb to ask, but in the docs of MPI.jl I/O I see no reference to functions such as iwrite_shared or iwrite_ordered for doing unblocking writes (and reads). Any reason for such choice?

P.S. In the docs of MPI.File.write_shared you say to look for MPI.File.write_ordered for the noncollective version. It's wrong. MPI.File.write_ordered is collective.

[vchuravy]

Welcome @carlocastoldi, if something is missing in MPI.jl it is often that no-one else had a particular need for it.
We welcome contributions and it shouldn't be to hard to add the necessary functions in https://github.com/JuliaParallel/MPI.jl/blob/master/src/io.jl