CUDA.md

CUDA-based MAX docker stack

GPU accelerated, multi-arch (linux/amd64, linux/arm64/v8) docker images:

• glcr.b-data.ch/cuda/max/base
• glcr.b-data.ch/cuda/max/scipy

Images available for MAX versions ≥ 24.6.0.

Build chain

The same as the MAX/Mojo docker stack.

Features

The same as the MAX/Mojo docker stack plus the CUDA runtime.

Table of Contents

• Prerequisites
• Install
• Usage

Prerequisites

The same as the MAX/Mojo docker stack plus

• NVIDIA GPU
• NVIDIA Linux driver
• NVIDIA Container Toolkit

ℹ️ The host running the GPU accelerated images only requires the NVIDIA driver, the CUDA toolkit does not have to be installed.

Install

To install the NVIDIA Container Toolkit, follow the instructions for your platform:

• Installation Guide — NVIDIA Cloud Native Technologies documentation

Usage

Build image (base)

latest:

docker build \
  --build-arg BASE_IMAGE=ubuntu \
  --build-arg BASE_IMAGE_TAG=22.04 \
  --build-arg BUILD_ON_IMAGE=glcr.b-data.ch/cuda/python/ver \
  --build-arg MOJO_VERSION=25.1.0 \
  --build-arg PYTHON_VERSION=3.12.9 \
  --build-arg CUDA_IMAGE_FLAVOR=base \
  --build-arg INSTALL_MAX=1 \
  -t cuda/max/base \
  -f base/latest.Dockerfile .

version:

docker build \
  --build-arg BASE_IMAGE=ubuntu \
  --build-arg BASE_IMAGE_TAG=22.04 \
  --build-arg BUILD_ON_IMAGE=glcr.b-data.ch/cuda/python/ver \
  --build-arg CUDA_IMAGE_FLAVOR=base \
  --build-arg INSTALL_MAX=1 \
  -t cuda/max/base:MAJOR.MINOR.PATCH \
  -f base/MAJOR.MINOR.PATCH.Dockerfile .

For MAJOR.MINOR.PATCH ≥ 24.6.0.

Run container

self built:

docker run -it --rm \
  --gpus '"device=all"' \
  cuda/max/base[:MAJOR.MINOR.PATCH]

from the project's GitLab Container Registries:

docker run -it --rm \
  --gpus '"device=all"' \
  IMAGE[:MAJOR[.MINOR[.PATCH]]]

IMAGE being one of

• glcr.b-data.ch/cuda/max/base
• glcr.b-data.ch/cuda/max/scipy

The CUDA-based MAX docker stack is derived from the CUDA-based Python docker stack.
ℹ️ See also Python docker stack > Notes on CUDA.