GPU Setup Guide

Configure your graphics card for optimal Enchant performance.

NVIDIA GPU Setup

NVIDIA GPUs deliver the best performance with Enchant through CUDA acceleration.

1. Update GPU Drivers

Download the latest Game Ready or Studio drivers from the NVIDIA Driver Downloads page. Studio drivers are recommended for production workstations.

# Check your current driver version
nvidia-smi

You should see output showing your GPU model, driver version, and CUDA version. The driver version should be 525.0 or newer.

2. CUDA Toolkit

Enchant bundles its own CUDA runtime, so you generally do not need to install the CUDA Toolkit separately. However, if you encounter CUDA-related errors:

# Install CUDA Toolkit 12.x (optional)
# Download from: https://developer.nvidia.com/cuda-downloads

# Verify installation
nvcc --version

3. cuDNN (Optional)

cuDNN is bundled with Enchant. Manual installation is only needed if you see specific cuDNN errors in the application logs.

# If needed, download cuDNN from:
# https://developer.nvidia.com/cudnn

# Copy files to CUDA directory:
copy cudnn\bin\*.dll "C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v12.x\bin"
copy cudnn\include\*.h "C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v12.x\include"
copy cudnn\lib\x64\*.lib "C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v12.x\lib\x64"

4. Verify Setup

# Confirm GPU is detected and CUDA works
nvidia-smi

# Expected output:
# +-----------------------------------------------------------------------------+
# | NVIDIA-SMI 545.xx    Driver Version: 545.xx    CUDA Version: 12.x          |
# |-------------------------------+----------------------+---------------------+
# | GPU  Name         | Bus-Id   | Disp.A | Volatile Uncorr. ECC |
# | ...               | ...      | ...    | ...                  |
# +-------------------------------+----------------------+---------------------+

AMD GPU Setup

AMD GPU support uses DirectML on Windows and ROCm on Linux.

Windows (DirectML)

On Windows, Enchant uses DirectML for AMD GPU acceleration. No additional setup is required beyond installing the latest AMD drivers:

1. Download the latest AMD Adrenalin drivers.
2. Run the installer and select "Recommended" setup.
3. Restart your computer.
4. Launch Enchant — AMD GPU should be detected automatically.

Linux (ROCm)

# Ubuntu 22.04
sudo apt update
sudo apt install rocm-dev rocm-libs

# Add user to render/video groups
sudo usermod -aG render,video $USER

# Verify
rocm-smi

Apple Silicon (M1/M2/M3)

Apple Silicon Macs use Metal Performance Shaders (MPS) for GPU acceleration.

Recommended macOS Versions

Performance Tips

• Close other GPU-intensive apps while running Enchant to maximize available VRAM.
• Use an SSD for your footage — disk I/O can be a bottleneck for large files.
• Keep drivers updated — newer drivers often include performance improvements for AI workloads.
• Monitor GPU temperature — sustained heavy inference can cause thermal throttling on laptops.

Troubleshooting

GPU Not Detected

• Ensure your GPU meets the minimum requirements (4 GB VRAM).
• Update to the latest GPU drivers.
• On laptops with dual GPUs (integrated + dedicated), make sure Enchant is set to use the dedicated GPU in your system's graphics settings.

VRAM Out of Memory

• Close other applications using GPU memory (web browsers, games, other editors).
• Reduce the batch size in Enchant settings (Settings → Performance → Batch Size).
• If your GPU has only 4 GB VRAM, use "Low VRAM Mode" in settings.

Driver Conflicts

• Uninstall old GPU drivers completely using DDU (Display Driver Uninstaller) before installing new ones.
• Avoid installing multiple CUDA toolkit versions simultaneously.
• If you use Docker or WSL2, ensure GPU passthrough is configured correctly.

Slow Performance

# Check GPU utilization while Enchant is running
nvidia-smi -l 1

# Look for:
# - GPU-Util should be 80-100% during analysis
# - Memory usage should not be at maximum
# - Temperature should be below throttle threshold