pycuda

pycuda

CONTAINERS IMAGES RUN BUILD

CONTAINERS

pycuda
Builds
Requires	L4T >=32.6
Dependencies	build-essential cuda python numpy
Dependants	l4t-diffusion l4t-ml l4t-pytorch l4t-tensorflow:tf1 l4t-tensorflow:tf2 stable-diffusion-webui
Dockerfile	Dockerfile
Images	dustynv/pycuda:r32.7.1 (2023-12-06, 0.4GB) dustynv/pycuda:r35.2.1 (2023-09-07, 5.0GB) dustynv/pycuda:r35.3.1 (2023-08-29, 5.0GB) dustynv/pycuda:r35.4.1 (2023-12-06, 5.0GB) dustynv/pycuda:r36.2.0 (2023-12-06, 3.5GB)

CONTAINER IMAGES

Repository/Tag	Date	Arch	Size
dustynv/pycuda:r32.7.1	2023-12-06	arm64	0.4GB
dustynv/pycuda:r35.2.1	2023-09-07	arm64	5.0GB
dustynv/pycuda:r35.3.1	2023-08-29	arm64	5.0GB
dustynv/pycuda:r35.4.1	2023-12-06	arm64	5.0GB
dustynv/pycuda:r36.2.0	2023-12-06	arm64	3.5GB

_{Container images are compatible with other minor versions of JetPack/L4T:
    • L4T R32.7 containers can run on other versions of L4T R32.7 (JetPack 4.6+)
    • L4T R35.x containers can run on other versions of L4T R35.x (JetPack 5.1+)}

RUN CONTAINER

To start the container, you can use the run.sh/autotag helpers or manually put together a docker run command:

# automatically pull or build a compatible container image
./run.sh $(./autotag pycuda)

# or explicitly specify one of the container images above
./run.sh dustynv/pycuda:r32.7.1

# or if using 'docker run' (specify image and mounts/ect)
sudo docker run --runtime nvidia -it --rm --network=host dustynv/pycuda:r32.7.1

^{run.sh forwards arguments to docker run with some defaults added (like --runtime nvidia, mounts a /data cache, and detects devices)
autotag finds a container image that's compatible with your version of JetPack/L4T - either locally, pulled from a registry, or by building it.}

To mount your own directories into the container, use the -v or --volume flags:

./run.sh -v /path/on/host:/path/in/container $(./autotag pycuda)

To launch the container running a command, as opposed to an interactive shell:

./run.sh $(./autotag pycuda) my_app --abc xyz

You can pass any options to run.sh that you would to docker run, and it'll print out the full command that it constructs before executing it.

BUILD CONTAINER

If you use autotag as shown above, it'll ask to build the container for you if needed. To manually build it, first do the system setup, then run:

./build.sh pycuda

The dependencies from above will be built into the container, and it'll be tested during. See ./build.sh --help for build options.