Projects using pkgpl_base docker images.
- docker for cpu.
- nvidia-docker for gpu.
In docker
directory, add additional packages you need to Dockerfile
.
Set PROJECT_NAME
and DEVICE
in Env.sh
and run following commands.
./build.sh # docker build
./run.sh # docker run
./exec.sh -h # docker exec
./helper.sh # docker images, ps, stop, rmi
Directory structure from Good enough practices in scientific computing.
bin
: put external scripts or compiled programsdata
: put raw data and metadatadoc
: put text documents associated with the projectresults
: put files generated during cleanup and analysissrc
: put project source code
Directories above will be mounted in the docker container.
- path:/home/$USER/$RPOJECT_DIR/
- link:/work/$USER/$PROJECT_DIR/