View a markdown version of this page

Build a Blender Docker image for GPU rendering on Deadline Cloud - Deadline Cloud

Build a Blender Docker image for GPU rendering on Deadline Cloud

The blender-aswf-ci-base example builds a Docker image that packages Blender with the deadline-cloud-for-blender adaptor and GPU support (CUDA/OptiX) for rendering on Deadline Cloud service-managed fleets. The base image is aswf/ci-base:2026 (Rocky Linux 8, CUDA 12.9, VFX Platform 2026).

You can use this image for the following purposes:

  • Run Blender Cycles GPU renders on Deadline Cloud service-managed fleets.

  • Bundle third-party Blender addons into the image at build time by placing addon .zip files in the plugins/ directory before building.

Build the image:

docker build -t blender-aswf:4.5.0 .

Push to Amazon Elastic Container Registry (Amazon ECR) and deploy the full stack (queue, fleet, and queue environment) with the included AWS CloudFormation (CloudFormation) template:

aws cloudformation deploy \
    --template-file cloudformation.yaml \
    --stack-name blender-aswf-ci-base-stack \
    --parameter-overrides \
        FarmId=farm-... \
        ECRImageURI=account-id.dkr.ecr.region.amazonaws.com/repo:4.5.0 \
        FleetRoleArn=arn:aws:iam::account-id:role/FleetRole \
        QueueRoleArn=arn:aws:iam::account-id:role/QueueRole \
        JobAttachmentsBucket=my-deadline-bucket

The queue role needs Amazon ECR pull permissions (ecr:BatchGetImage, ecr:GetDownloadUrlForLayer, and ecr:GetAuthorizationToken) to pull the container image at runtime.

GPU rendering is automatic when the fleet has GPU instances. The queue environment conditionally adds --gpus all --runtime=nvidia based on whether the host has an NVIDIA GPU. CPU-only instances fall back to Cycles CPU rendering.

For a job bundle that renders Blender scenes, see Render Blender scenes on Deadline Cloud.