Docker recipies for TCLB, hosted at https://hub.docker.com/u/mdzik
- cuda in image name indicate GPU support
- core images are not intendet for direct use
- buildkit images should suffice to build full-futered TCLB
- workspace CPU only are intendet for workshops, they include jupyter bindings. See also mdzik/TCLB_binder repo
- install Singularity (https://sylabs.io/guides/3.8/user-guide/quick_start.html#quick-installation-steps)
- clone this repo
git clone https://github.com/mdzik/TCLB_docker.git
cd TCLB_dockernow you need to set path to your TCLB clone repo (cloned fork probably)
make clean
make activate
editor .local/config_alland you are good to go
source activate workspace_cpu
cd $TCLB_PATH
scmd ./configure $CONFIGUREARGS
tclbmake -j8 d2q9
tclb d2q9 $TCLB_PATH/example/flow/2d/karman.xml - Keep in mind that Singularity will include in your env directories below PWD, not necessary home.
- Jupyter Lab will do the same, so
cdbeforescmd/jupyterlab