BRaaS-HPC-Interactive is a Blender addon that enables interactive rendering and computational work on High-Performance Computing (HPC) clusters. This addon extends the base BRaaS-HPC functionality by providing real-time, interactive sessions with HPC resources, allowing users to run custom commands and scripts directly on HPC nodes while working within Blender.
Developed by IT4Innovations National Supercomputing Center, this addon is designed for users who need to leverage HPC computational power for:
- Interactive Blender rendering (BlenderPhi)
- Real-time visualization and processing
- Remote computational tasks with live feedback
- Live HPC Sessions: Establish interactive SSH connections to HPC compute nodes
- Interactive Types: Currently only supports BlenderPhi
- Real-time Command Execution: Run commands and scripts on HPC resources with live output
- Automatic SSH Tunneling: Secure port forwarding between local machine and HPC nodes
This add-on currently supports one HPC system (more systems will be added):
- KAROLINA (IT4Innovations)
- Submit interactive jobs to HPC clusters
- Monitor job status and execution
- File transfer between local machine and HPC storage
- Automatic resource allocation (CPU/GPU)
- Automatic installation of required scripts on HPC clusters
- BlenderPhi installation and configuration
- Git-based script repository management
- Cluster-specific script configurations
- Blender: 4.5.0 or higher
- Operating System: Windows, Linux, or macOS
- Python: 3.x (bundled with Blender)
- SSH Client: OpenSSH or compatible SSH client
- Dependencies:
- SSH access with key-based authentication
- SLURM or PBS job scheduler
- Shared storage accessible from compute nodes
- Git installed (for script installation)
- BlenderPhi installed (or can be installed via addon)
Download the add-on in zip format: https://github.com/It4innovations/braas-hpc-interactive/releases
- Open Blender (version 4.5.0 or higher)
- Go to
Edit > Preferences > Add-ons - Click
Install...button - Navigate to the
braas_hpc_interactive.zipfile and install it - Enable the add-ons by checking the checkboxes next to "BRaaS-HPC", "BRaaS-HPC-Interactive" and "BRaaS-HPC-RenderEngine".
- In Blender, go to
Edit > Preferences > Add-ons - Find "BRaaS-HPC-Interactive" addon
- Expand the addon preferences
- Configure the following:
- Git Repository (Scripts): Default is
https://github.com/It4innovations/braas-hpc-interactive.git - Branch: Default is
main - Link (BlenderPhi): URL to BlenderPhi download for HPC (e.g.,
https://code.it4i.cz/raas/blenderphi/-/blob/main/releases/blenderphi-v4.5.5/blenderphi-v4.5.5-karolina-linux-x64-gcc13.tar.xz)
- Git Repository (Scripts): Default is
- Click
Install scripts on the cluster(s)button - Wait for confirmation that scripts are installed
-
Configure Cluster Settings (in base BRaaS-HPC addon):
- Open addon preferences for the base
braas_hpcaddon - Add cluster configurations (hostname, username, SSH key path, project/allocation)
- Enable the clusters you want to use
- Open addon preferences for the base
-
Select Configuration (in Blender scene):
- Go to the Blender's Render Properties panel
- Look for the "BRaaS-HPC" panel
- Select your cluster, partition, and allocation
-
Open Interactive Panel:
- In the Blender's Render Properties panel, find the "BRaaS-HPC Interactive" section under "BRaaS-HPC"
-
Select Interactive Type:
- Choose from dropdown:
BlenderPhi(default) - (More interactive types will be added)
- Choose from dropdown:
-
Submit Job:
- Click
Submit Interactive Job - The addon will:
- Package your Blender file
- Upload it to the HPC cluster
- Submit a job to the scheduler
- Wait for resources to be allocated
- Click
-
Monitor Job Status:
- Refresh the job list to see when your job is running
- Once in "RUNNING" state, you can connect to it
-
Execute Command:
- Click
Run Interactive Command - The addon will:
- Establish SSH tunnel to the compute node
- Execute the command in the background
- Keep the connection alive for monitoring
- Click
-
Stop Command:
- Click
Stop Interactive Commandto terminate the running process - This closes the SSH tunnel and stops remote execution
- Click
- Local Storage: Files are stored in the path configured in BRaaS-HPC preferences
- Remote Storage: Files are uploaded to your cluster's shared storage
- Automatic Transfer: The addon handles file transfers automatically
- Job Structure:
<job_name>/ ├── in/ # Input files (Blender file, assets) ├── out/ # Output files (renders, logs) └── interactive/ # Interactive session files ├── blendfile # Path to Blender file ├── hostname.txt # Compute node hostname ├── jobid.txt # Job ID ├── command.sh # Commands to execute ├── command.log # Standard output └── command.err # Error output
1. Configure Cluster (Preferences)
↓
2. Select Cluster/Partition/Allocation (Viewport Panel)
↓
3. Choose Interactive Type (Interactive Panel)
↓
4. Submit Interactive Job (Button)
↓
5. Wait for Job to Start (Monitor in Job List)
↓
6. Run Interactive Command (Button)
↓
7. Work Interactively / Monitor Output
↓
8. Stop Command When Done (Button)
↓
9. Download Results (if needed)
- SSH Key Authentication: Secure key-based authentication (no passwords stored)
- Strict Host Checking: Can be configured per security requirements
- Encrypted Tunnels: All data transfers through encrypted SSH connections
-
SSH Connection Fails:
- Verify SSH key is correctly installed on HPC cluster
- Check firewall settings allow SSH (port 22)
- Ensure correct username and hostname in cluster configuration
- Test manual SSH connection:
ssh -i /path/to/key user@host
-
Script Installation Fails:
- Check internet connectivity from HPC cluster
- Verify Git is installed on cluster
- Ensure write permissions in home directory
- Check repository URL and branch name
-
Job Submission Fails:
- Verify project/allocation is active and has resources
- Check job parameters (nodes, walltime, partition)
- Review cluster-specific requirements
- Check HPC scheduler status
-
Interactive Command Doesn't Start:
- Ensure job is in "RUNNING" state
- Check if BlenderPhi is installed on cluster
- Verify port forwarding settings
- Review command logs on cluster (
command.log,command.err)
-
File Transfer Issues:
- Check storage path permissions
- Verify sufficient disk space on both local and remote
- Ensure network connectivity is stable
- Check file paths don't contain special characters
This software is licensed under the terms of the GNU General Public License.
This work was supported by the Ministry of Education, Youth and Sports of the Czech Republic through the e-INFRA CZ (ID:90254).
This work was supported by the SPACE project. This project has received funding from the European High- Performance Computing Joint Undertaking (JU) under grant agreement No 101093441. This project has received funding from the Ministry of Education, Youth and Sports of the Czech Republic (ID: MC2304).