General Information

Blender is a free, open‑source 3D creation suite used for modeling, animation, rendering, visual effects, video editing, game asset creation, and more. It is developed by the Blender Foundation and supported by a large global community. It's completely free and open source under the GPL license, with full access to its source code.

Running Blender

Discover available Blender versions

Discover available Blender versions
ml spider blender

Slurm Job Templates

Blender detects automatically how many CPUs and GPUs are available without oversubscription.

Only the `CYCLES` rendering engine works non-interactively. EEVEE won't work.

While you can in principle run Blender via X-Forwarding, but it is incredibly slow. Every action takes several seconds. Better avoid by preparing blender inputs locally.


Example Sbatch Template (CPU-only) 
#!/usr/bin/env bash

# Job name
#SBATCH --job-name=blender-example

# Select a partition
#SBATCH --partition=epyc

#Select number of nodes
#SBATCH --nodes=1

# Select number of tasks per node (Blender binaries don't support MPI)
#SBATCH --tasks-per-node=1

# Select number of threads per task (OpenMP)
#SBATCH --cpus-per-task=16

# Request memory (scale up dynamically)
#SBATCH --mem-per-cpu=4G

# Timelimit
#SBATCH --time=1-0

# Load required modules
ml purge
ml blender

# Run blender non-interactively
blender -b -E CYCLES -o "//output/blender-test-###.png" -a slample.blend


If you are tempted to use GPUs, have at least the courtesy to compare against a CPU-only setup!


Example Sbatch Template (CPU+GPU) 
#!/usr/bin/env bash

# Job name
#SBATCH --job-name=blender-example

# Select a partition
#SBATCH --partition=epyc

#Select number of nodes
#SBATCH --nodes=1

# Select number of tasks per node (Blender binaries don't support MPI)
#SBATCH --tasks-per-node=1

# Select number of threads per task (OpenMP)
#SBATCH --cpus-per-task=16

# Request memory (scale up dynamically)
#SBATCH --mem-per-cpu=4G

# Request a GPU
#SBATCH --gpus=1

# Timelimit
#SBATCH --time=1-0

# Load required modules
ml purge
ml cuda blender

# Run blender non-interactively
blender -b -E CYCLES -o "//output/blender-test-###.png" -a slample.blend -- --cycles-device CUDA+CPU


Support

If you have any problems with Blender please contact the HPC-Servicedesk.

Also, if you have improvements to this documentation that other users can profit from, please reach out!

  • Keine Stichwörter

part of icons are from social media marketing icons Social media marketing icons created by Freepik - Flaticon