The programming support for NVIDIA GPUs in Julia is provided by the
CUDA.jl package. It does not require the
entire CUDA toolkit installed, as it will automatically be downloaded when the
package is first used.
ℹ️ This docker stack is derived from a CUDA image since Python
packages may require the CUDA toolkit.
These images are tweaked as follows:
The following startup scripts are put in place:
- $JULIA_PATH/etc/julia/startup.jl to add the
LOAD_PATHof the pre-installed packages - $HOME/.julia/config/startup.jl to start Revise and activate either the project environment or package directory.
Versions
JULIA_VERSIONPYTHON_VERSIONGIT_VERSIONGIT_LFS_VERSIONPANDOC_VERSIONQUARTO_VERSION(pubtools image)
Miscellaneous
BASE_IMAGE: Its very base, a Docker Official Image.PARENT_IMAGE: The image it was derived from.BUILD_DATE: The date it was built (ISO 8601 format).CTAN_REPO: The CTAN mirror URL. (pubtools image)
In addition to the TeX packages used in
rocker/verse,
jupyter/scipy-notebook
and required for nbconvert, the
packages requested by the community
are installed.
The Python version is selected as follows:
- The latest Python version numba is compatible with.
This Python version is installed at /usr/local/bin.
The CUDA and OS versions are selected as follows:
- CUDA: The lastest version that has image flavour
develincluding cuDNN available. - OS: The latest version that has TensortRT libraries for
amd64available.
ℹ️ It is taking quite a long time for these to be available forarm64.
Versions
CUDA_VERSION
Miscellaneous
CUDA_IMAGE: The CUDA image it is derived from.