An app to demo nirs mcx simulations.

• a compatible nvidia gpu
  • compatible meaning, at least compute-capability 3.0, see list of here
  • nvidia driver installed
• docker
• nvidia-docker
• to configure the simulations, experience with mcx, Monte Carlo eXtreme, is needed.
• understanding of chrisapp would be useful, but not necessary

1. incoming: input directory for files.
2. outgoing: output directory for files.
3. head: NIfTI file containing head to simulate.
4. mcx_config: json file containing options for how to run mcx, and vectors for srcpos, srcdir, detpos, and detdir.

{
    "mcx" : {
        "gpuid": "1",
        "prop": [
            [0, 0, 1, 1],
            [0.014600,  8.28, 0.9, 1.37],
            [0.019500, 11.11, 0.9, 1.37]
        ]
    },
    "srcpos": [53, 100, 50],
    "srcdir": [0.9063077870366499, 0, 0.42261826174069944],
    "detpos": [47, 100,  62.5],
    "detdir": [0.9063077870366499, 0, 0.42261826174069944]
}

Using the example config and head, here is an example of how to run this plugin using nvidia-docker.

mkdir -p out && nvidia-docker run -v example:/incoming -v out:/outgoing   \
             jdtatz/pl-nirs-demo-app nirs_demo_app.py /incoming /outgoing \
             --head=example_head.nii.gz --mcx_config=example_config.json