Support MPI.PROC_NULL in send/receive (plot_merged)

[carstenbauer]

I ran this code, which produced tapes for, say, 10 MPI ranks.

using MPI
using MPITape

function main()
    MPI.Init()

    comm = MPI.COMM_WORLD
    rank = MPI.Comm_rank(comm)
    nranks = MPI.Comm_size(comm)

    # chain topology, i.e. "open" boundary conditions
    left  = rank - 1
    right = rank + 1
    if left < 0
        left = MPI.PROC_NULL
    elseif right >= nranks
        right = MPI.PROC_NULL
    end

    N = 2^22
    buf = rand(N)

    for r in 0:nranks-1
        MPI.Barrier(comm)
        if r == rank
            rightstr = right == MPI.PROC_NULL ? "no one" : right
            leftstr = left == MPI.PROC_NULL ? "no one" : left
            println("Rank $r will send to $rightstr and receive from $leftstr soon.")
        end
    end

    # this communication will happen in serial
    MPI.Send(buf, comm; dest=right)
    MPI.Recv!(buf, comm; source=left)

    # Due to serialization, ranks will reach this point in a deterministic order.
    # The last rank will reach this point first, then the second to last rank, and so on.
    println("Rank $rank has finished Recv!/Send.")

    MPI.Finalize()
end

@record main()
MPITape.save()

Afterwards I ran

tape_merged = MPITape.readall_and_merge()
MPITape.print_merged(tape_merged)

to get

[...]
    # 7: Send -> 8           (Δt=5.54E+01)
    # 8: Send -> 9           (Δt=5.55E+01)
    # 2: Send -> 3           (Δt=5.55E+01)
    # 5: Send -> 6           (Δt=5.55E+01)
    # 1: Send -> 2           (Δt=5.55E+01)
    # 6: Send -> 7           (Δt=5.55E+01)
    # 3: Send -> 4           (Δt=5.55E+01)
    # 4: Send -> 5           (Δt=5.55E+01)
    # 0: Send -> 1           (Δt=5.55E+01)
    # 9: Send -> -1          (Δt=5.70E+01)
    # 9: Recv <- 8           (Δt=5.71E+01)
    # 8: Recv <- 7           (Δt=5.73E+01)
    # 7: Recv <- 6           (Δt=5.74E+01)
    # 6: Recv <- 5           (Δt=5.76E+01)
    # 5: Recv <- 4           (Δt=5.77E+01)
    # 4: Recv <- 3           (Δt=5.79E+01)
    # 3: Recv <- 2           (Δt=5.80E+01)
    # 2: Recv <- 1           (Δt=5.82E+01)
    # 1: Recv <- 0           (Δt=5.83E+01)
    # 0: Recv <- -1          (Δt=5.85E+01)

Note the -1 which corresponds to MPI.PROC_NULL (i.e. a dummy send/receive).

Trying to run MPITape.plot_merged(tape_merged) I got

julia> MPITape.plot_merged(tape_merged)
ERROR: Not all destinations found for MPITape.MPIEvent(9, MPI_Send, ..., 57.006352186203, 57.00636911392212): [-1]
Stacktrace:
 [1] error(s::String)
   @ Base ./error.jl:35
 [2] get_edges(tape::Vector{MPITape.MPIEvent}; check::Bool)
   @ MPITape /scratch/pc2-mitarbeiter/bauerc/.julia/packages/MPITape/FbfeO/src/communication_graph.jl:99
 [3] get_edges
   @ /scratch/pc2-mitarbeiter/bauerc/.julia/packages/MPITape/FbfeO/src/communication_graph.jl:62 [inlined]
 [4] plot_merged(tape::Vector{MPITape.MPIEvent}; palette::PlotUtils.ColorPalette, fname::String)
   @ MPITape /scratch/pc2-mitarbeiter/bauerc/.julia/packages/MPITape/FbfeO/src/plotting.jl:36
 [5] plot_merged(tape::Vector{MPITape.MPIEvent})
   @ MPITape /scratch/pc2-mitarbeiter/bauerc/.julia/packages/MPITape/FbfeO/src/plotting.jl:25
 [6] top-level scope
   @ REPL[4]:1

We should try to support -1 here in the sense of not showing an arrow for the send/recv.

cc @Mellich