MPI is based on a single program, multiple data (SPMD) model, where multiple processes are launched running independent programs, which then communicate as necessary via messages.
A script should include
using MPI and
MPI.Init() statements, for example
# examples/01-hello.jl using MPI MPI.Init() comm = MPI.COMM_WORLD println("Hello world, I am $(MPI.Comm_rank(comm)) of $(MPI.Comm_size(comm))") MPI.Barrier(comm)
The program can then be launched via an MPI launch command (typically
$ mpiexec -n 3 julia --project examples/01-hello.jl Hello world, I am rank 0 of 3 Hello world, I am rank 2 of 3 Hello world, I am rank 1 of 3
CUDA-aware MPI support
If your MPI implementation has been compiled with CUDA support, then
CuArrays (from the CuArrays.jl package) can be passed directly as send and receive buffers for point-to-point and collective operations (they may also work with one-sided operations, but these are not often supported).
In order to ensure MPI routines are called in the correct order at finalization time, MPI.jl maintains a reference count. If you define an object that needs to call an MPI routine during its finalization, you should call
MPI.refcount_inc() when it is initialized, and
MPI.refcount_dec() in its finalizer (after the relevant MPI call).
mutable struct MyObject ... function MyObject(args...) obj = new(args...) # MPI call to create object refcount_inc() finalizer(obj) do x # MPI call to free object refcount_dec() end return obj end end