DOCS: Add GCP Vertex AI docs

Dockerfile

@@ -0,0 +1,92 @@
+# Maybe have to transfer from Debian
+ARG RAPIDS_IMAGE
+FROM $RAPIDS_IMAGE as rapids
+
+RUN conda list -n rapids --explicit > /rapids/rapids-spec.txt
+
+FROM gcr.io/deeplearning-platform-release/rapids-gpu.21-12
+
+COPY --from=rapids /rapids/rapids-spec.txt /tmp/spec.txt
+
+RUN conda create --name rapids --file /tmp/spec.txt && \
+    rm -f /tmp/spec.txt
+
+ENV CONDA_DEFAULT_ENV=rapids
+
+#CMD python -c "import platform; print(platform.python_version())"
+# 3.7.12
+
+#CMD python -c "import platform; print(platform.platform())"
+# Linux-5.10.109-0-virt-x86_64-with-debian-bullseye-sid
+
+#CMD printenv
+# GDAL_DATA=/opt/conda/share/gdal
+# NV_LIBNPP_DEV_VERSION=11.1.0.245-1
+# CUDA_PATH=/opt/conda
+# NVIDIA_VISIBLE_DEVICES=all
+# NCCL_VERSION=2.12.10-1
+# CONDA_PREFIX=/opt/conda
+# GSETTINGS_SCHEMA_DIR_CONDA_BACKUP=
+# CONDA_EXE=/opt/conda/bin/conda
+# JAVA_HOME=/opt/conda
+# PWD=/
+# JAVA_LD_LIBRARY_PATH_BACKUP=
+# PROJ_LIB=/opt/conda/share/proj
+# LC_ALL=C.UTF-8
+# NV_CUDNN_VERSION=8.0.5.39
+# NV_NVTX_VERSION=11.0.167-1
+# NV_LIBNPP_VERSION=11.1.0.245-1
+# NV_LIBNCCL_DEV_PACKAGE=libnccl-dev=2.12.10-1+cuda11.0
+# CONDA_DEFAULT_ENV=base
+# SHELL=/bin/bash
+# NV_LIBCUBLAS_DEV_PACKAGE=libcublas-dev-11-0=11.2.0.252-1
+# GSETTINGS_SCHEMA_DIR=/opt/conda/share/glib-2.0/schemas
+# CONDA_PYTHON_EXE=/opt/conda/bin/python
+# NV_CUDA_CUDART_DEV_VERSION=11.0.221-1
+# CONTAINER_NAME=rapids-gpu/21-12+cu110
+# LANG=C.UTF-8
+# NV_LIBNCCL_DEV_PACKAGE_NAME=libnccl-dev
+# NV_LIBCUSPARSE_DEV_VERSION=11.1.1.245-1
+# NV_LIBNCCL_PACKAGE=libnccl2=2.12.10-1+cuda11.0
+# DL_ANACONDA_HOME=/opt/conda
+# NV_LIBCUBLAS_DEV_PACKAGE_NAME=libcublas-dev-11-0
+# NV_LIBCUBLAS_PACKAGE=libcublas-11-0=11.2.0.252-1
+# NVARCH=x86_64
+# PROJ_NETWORK=ON
+# PATH=/opt/conda/bin:/opt/conda/condabin:/opt/conda/bin:/usr/local/nvidia/bin:/usr/local/cuda/bin:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin
+# NV_CUDA_CUDART_VERSION=11.0.221-1
+# ANACONDA_PYTHON_VERSION=3.7
+# _CE_CONDA=
+# NV_CUDNN_PACKAGE_DEV=libcudnn8-dev=8.0.5.39-1+cuda11.0
+# NV_NVML_DEV_VERSION=11.0.167-1
+# NV_LIBNPP_DEV_PACKAGE=libnpp-dev-11-0=11.1.0.245-1
+# CONTAINER_URL=us-docker.pkg.dev/deeplearning-platform-release/gcr.io/rapids-gpu.21-12:nightly-2022-05-26
+# TERM=xterm
+# NV_LIBCUSPARSE_VERSION=11.1.1.245-1
+# NVIDIA_DRIVER_CAPABILITIES=compute,utility
+# NV_CUDA_LIB_VERSION=11.0.3-1
+# NV_LIBNCCL_PACKAGE_NAME=libnccl2
+# NVIDIA_REQUIRE_CUDA=cuda>=11.0 brand=tesla,driver>=418,driver<419
+# NV_LIBCUBLAS_PACKAGE_NAME=libcublas-11-0
+# NV_NVPROF_VERSION=11.0.221-1
+# NV_CUDNN_PACKAGE=libcudnn8=8.0.5.39-1+cuda11.0
+# CUDA_VERSION=11.0.3
+# _CE_M=
+# NV_LIBNPP_PACKAGE=libnpp-11-0=11.1.0.245-1
+# NV_LIBNCCL_DEV_PACKAGE_VERSION=2.12.10-1
+# CPL_ZIP_ENCODING=UTF-8
+# NV_CUDNN_PACKAGE_NAME=libcudnn8
+# JAVA_LD_LIBRARY_PATH=/opt/conda/lib/server
+# NV_LIBCUBLAS_DEV_VERSION=11.2.0.252-1
+# HOME=/root
+# CONDA_SHLVL=1
+# SHLVL=0
+# LD_LIBRARY_PATH=/usr/local/cuda/lib64:/usr/local/cuda/lib:/usr/local/lib/x86_64-linux-gnu:/usr/local/nvidia/lib:/usr/local/nvidia/lib64:/usr/local/nvidia/lib:/usr/local/nvidia/lib64
+# NV_LIBNCCL_PACKAGE_VERSION=2.12.10-1
+# HOSTNAME=6b8a39109ad6
+# JAVA_HOME_CONDA_BACKUP=
+# NV_LIBCUBLAS_VERSION=11.2.0.252-1
+# NV_NVPROF_DEV_PACKAGE=cuda-nvprof-11-0=11.0.221-1
+# CONDA_PROMPT_MODIFIER=(base)
+# NV_CUDA_COMPAT_PACKAGE=cuda-compat-11-0
+# LIBRARY_PATH=/usr/local/cuda/lib64/stubs

source/cloud/gcp/index.md

@@ -6,4 +6,5 @@ maxdepth: 2
 caption: Google Cloud Platform
 ---
 single-node
+vertex_ai
 ```

source/cloud/gcp/vertex_ai.md

@@ -0,0 +1,23 @@
+# Vertex AI
+
+RAPIDS can be deployed on Vertex AI.
+
+## Managed Notebooks
+
+While Vertex AI provides a `RAPIDS 0.18` environment for new, user-managed notebooks, it is recommended to a RAPIDS docker image to access the latest RAPIDS software.
+
+**0. Prepare RAPIDS Docker Image.** Before configuring a new notebook, the [RAPIDS Docker image](#rapids-docker) will need to be built to expose port 8080 to be used as a notebook service.
+
+```dockerfile
+ARG RAPIDS_IMAGE
+FROM $RAPIDS_IMAGE as rapids
+EXPOSE 8080
+
+ENTRYPOINT [ "jupyter-lab", "--allow-root", "--ip=0.0.0.0", "--port=8080", "--no-browser", "--NotebookApp.token=''", "--NotebookApp.allow_origin='*'" ]
+```
+
+Once you have built this image, it needs to be pushed to [Google Container Registry](https://cloud.google.com/container-registry/docs/pushing-and-pulling) for Vertex AI.
+
+**1. Create a New Notebook.** From the Google Cloud UI, nagivate to `Vertex AI` -> `Dashboard` and select `+ CREATE NOTEBOOK INSTANCE`. Under the `Environment` section, specity `Custom container`, and in the section below, select the `gcr.io` path to your pushed RAPIDS Docker image. After customizing other aspects of the machine, click `CREATE`
+
+**2. TEST RAPIDS** Once the managed notebook is fully configured, you can click `OPEN JUPYTERLAB` to navigate to another tab running JupyterLab to use the latest version of RAPIDS with Vertex AI.