Docker example - mpi

This example starts a container running two single-host MPI programs:

1. A simple hello world program, which you can inspect (and alter) in helloworld.c.

2. A simple matmult (matrix multiplication) program, which you can inspect (and alter) in matmult.c.

Building the image

# View available docker images
[~]: docker images

# Change into this directory
[~]: cd /path/to/docker-examples/mpi

# Build the docker image for this example (line breaks for readability)
[/path/to/docker-examples/mpi]: docker build
        --tag="docker-monash/mpi-img" \ # Tag this image with this name
        --rm=true ./ # remove intermediate containers after successful build

# Watch it build. It may take some time the first time you run it.

# Convince yourself the above image has been built
[/path/to/docker-examples/mpi]: docker images

Running the application

# These commands can be run from anywhere, so we assume you're in your home directory (~)
[~]: docker run \
    --name="singlenode.mpi" \
    --hostname="singlenode.mpi" \
    docker-monash/mpi-img

# Alternatively, you can run the example interactively WITHIN the container with:
[~]: docker run \
    --name="singlenode.mpi" \
    --hostname="singlenode.mpi" \
    -it \ # This tells docker to run the container interactively
    docker-monash/mpi-img \
    bash # this tells docker to run the bash when the container starts

    # Now you are INSIDE the container you just started. Kind of like SSH-ing into it
    [root@singlenode /]: cd docker-monash-mpi/
    # Run the example
    [root@singlenode docker-monash-mpi/]: ./run_mpi_example.sh
    # watch the example run
    # Exit the container and let it die.
    [root@singlenode docker-monash-mpi/]: exit

# Convince yourself the container has run
[~]: docker ps -a

# Now kill and remove the container
[~]: docker kill singlenode.mpi && docker rm singlenode.mpi

Notes

These containers will sip on your local computer's RAM and munch on CPU using Docker's defaults. You can, if you so wish, play around with docker run options --cpu-shares and --cpuset to limit how much of a CPU a container is allowed to use or which CPUs it can use respectively. You can also limit the memory it sips on using the --memory option. Check out the docs for more info and also here for a great example.