cc_staff
123
edits
RAPIDS: Difference between revisions
→Where to look for a Docker image for RAPIDS
| Line 17: | Line 17: | ||
There are two major sites where to look for a Docker image for RAPIDS. | There are two major sites where to look for a Docker image for RAPIDS. | ||
* [https://ngc.nvidia.com/catalog/containers/nvidia:rapidsai:rapidsai NVIDIA GPU Cloud (NGC)] | * [https://ngc.nvidia.com/catalog/containers/nvidia:rapidsai:rapidsai NVIDIA GPU Cloud (NGC)]: The RAPIDS Docker images are available on NGC in two types, i.e. | ||
** base - contains a RAPIDS environment ready to use. Use this image if you want to submit a job to the Slurm scheduler. | ** base type - contains a RAPIDS environment ready to use. Use this type of image if you want to submit a job to the Slurm scheduler. | ||
** runtime - extends the base image by adding a Jupyter notebook server and example notebooks. Use this | ** runtime type - extends the base image by adding a Jupyter notebook server and example notebooks. Use this type of images if you want to interactively work with RAPIDS through notebooks and examples. | ||
* Docker Hub | There are multiple images with different combinations of RAPIDS versions and CUDA versions with Ubuntu or CentOS base images for both types. You can find the Docker pull command of a selected image via the [https://ngc.nvidia.com/catalog/containers/nvidia:rapidsai:rapidsai/tags Tags] tab on that page. | ||
* [https://hub.docker.com/r/rapidsai/rapidsai-dev Docker Hub]: The RAPIDS Docker images on the rapidsai-dev repo on Docker Hub provides RAPIDS Docker images as devel type. | |||
** devel - contains the full RAPIDS source tree, the compiler toolchain, the debugging tools, the headers and the static libraries for RAPIDS development. Use this type of images if you want to implement any customized operations with low-level access to cuda-based processes. | |||
==Build a RAPIDS Singularity image== | ==Build a RAPIDS Singularity image== | ||