cuda-images's Introduction

`pyhf` NVIDIA CUDA enabled Docker images

pyhf Docker images built on the NVIDIA CUDA enabled images for runtime use with the the NVIDIA Container Toolkit.

These images are meant to be used as base images for bespoke user application images to build on.

Installation

Make sure that you have the nvidia-container-toolkit installed on the host machine
Check the list of available tags on Docker Hub to find the tag you want
Use docker pull to pull down the image corresponding to the tag

Example:

docker pull pyhf/cuda:0.6.3-jax-cuda-11.1

Use

To check that NVIDIA GPUS are being properly detected run

docker run --rm --gpus all pyhf/cuda:0.6.3-jax-cuda-11.1 'nvidia-smi'

and check if the nvidia-smi output appears correctly.

To run (interactively) using GPUs on the host machine:

docker run --rm -ti --gpus all pyhf/cuda:0.6.3-jax-cuda-11.1

Tests

To verify things are working on your host machine you can run

docker run --rm --gpus all -v $PWD:$PWD -w $PWD pyhf/cuda:0.6.3-jax-cuda-11.1 'bash tests/test_backend.sh jax'

cuda-images's People

Contributors

Stargazers

Watchers

cuda-images's Issues

Reduce image size with multi-stage builds

It may be possible to reduce the image sizes with multi-stage builds, though this might take lots of effort.

Some things that might be worth reading after a Google search on "NVIDIA multi stage build" are:

Smaller Docker Image using Multi-Stage Build
NVIDIA/nvidia-docker#675
A complete guide to building a Docker Image serving a Machine learning system in Production (section 3. Building a Docker image for any Python Project (GPU))

I'm not sure if mutli-stage builds help much though for Python projects. I think you would normally use the devel tagged nvidia/cuda images to build software and then once it is build copy it out in a secondary stage to a runtime tagged nvidia/cuda image.

Recommend Projects