🍓 StrawberryFist VRAM Optimizer for ComfyUI

A comprehensive VRAM management solution for ComfyUI that automatically optimizes GPU memory usage and provides real-time monitoring.

🎯 Features

🔧 VRAM Optimizer

• Automatic VRAM cleanup after each queue execution
• Multiple cleaning modes: Standard and Aggressive
• Smart cleaning conditions: Clean every time or only when memory usage is high
• Flexible timing: Clean before queue, after queue, or both
• Manual trigger: Force cleanup with a simple parameter change
• Real-time logging: Monitor all activities in the terminal

📊 GPU Monitor

• Real-time GPU monitoring with background thread
• Memory usage visualization with color-coded progress bars
• Historical data tracking with configurable history length
• Warning system with customizable thresholds
• Trend analysis showing memory usage patterns
• Multiple outputs for integration with other nodes

🚀 Installation

Method 1: ComfyUI Manager (Recommended)

1. Open ComfyUI Manager
2. Search for "StrawberryFist VRAM Optimizer"
3. Click Install
4. Restart ComfyUI

Method 2: Manual Installation

1. Navigate to your ComfyUI custom nodes directory:

   cd ComfyUI/custom_nodes/
   

2. Clone this repository:

   git clone https://github.com/YourUsername/comfyui-strawberryfist-vram-optimizer.git
   

3. Restart ComfyUI

Method 3: Download ZIP

1. Download the latest release from Releases

2. Extract to ComfyUI/custom_nodes/

3. Restart ComfyUI

4. Example

🔄 Automatic Execution Behavior The StFist - VRAM Optimizer node does not require a connection to other nodes in the workflow. Simply placing it anywhere on the canvas will trigger its execution automatically before and/or after each queue run, depending on the selected run_timing setting.

✨ Optimization status is:

Displayed directly in the ComfyUI console output (with timestamps and memory usage details)

Also optionally passed to the status output pin for use with nodes like Show Text

This allows for non-intrusive background optimization without disrupting the workflow structure.

✅ Tip: You can place the node in a corner of your workspace for passive VRAM management.

📋 Requirements

• ComfyUI
• Python 3.8+
• PyTorch with CUDA support
• GPUtil (automatically installed)

🎮 Usage

VRAM Optimizer Node

1. Add the StFist - VRAM Optimizer node to your workflow
2. Configure the settings:
   • enabled: Turn automatic cleaning on/off
   • clear_mode: Choose between Standard or Aggressive cleaning
   • auto_clean: Set cleaning conditions (Every Time or Only When High)
   • run_timing: Choose when to clean (After Queue, Before Queue, or Both)
   • force_run: Change this value to manually trigger cleaning

GPU Monitor Node

1. Add the StFist - GPU Monitor node to your workflow
2. Configure the monitoring settings:
   • monitoring_enabled: Turn real-time monitoring on/off
   • update_interval: Set monitoring frequency (0.1-10 seconds)
   • history_length: Number of data points to keep (10-300)
   • warning_threshold: Memory usage warning level (50-95%)
   • refresh_trigger: Change to force immediate update

📸 Screenshots

VRAM Optimizer in Action

🎮 ═══════════════════════════════════════════════════════════
🔥 StrawberryFist VRAM Optimizer v3.0 [14:23:45]
🔄 실행 횟수: 5 | 상태: CLEANED_1024.3MB
⚙️ 설정: After Queue / Standard
═══════════════════════════════════════════════════════════

📊 GPU 메모리 상태 (NVIDIA GeForce RTX 4090)
┌─────────────────────────────────────────────────────────┐
│ ✅ 사용률: 15.2% (GOOD)                                  │
│ 📈 사용량: 3724.1MB / 24564.0MB                         │
│ 🟢🟢🟢🟢⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜ │
└─────────────────────────────────────────────────────────┘

GPU Monitor Dashboard

🎮 ═══════════════════════════════════════════════════════════
🔥 StrawberryFist GPU 실시간 모니터링 [14:23:45]
═══════════════════════════════════════════════════════════

📊 GPU: NVIDIA GeForce RTX 4090
┌─────────────────────────────────────────────────────────┐
│ ⚠️ 사용률: 75.3% (WARNING)                               │
│ 📈 사용량: 18500.2MB / 24564.0MB                        │
│ 🟡🟡🟡🟡🟡🟡🟡🟡🟡🟡🟡🟡🟡🟡🟡🟡🟡🟡🟡⬜⬜⬜⬜⬜⬜ │
└─────────────────────────────────────────────────────────┘

📈 최근 통계 (최근 10개 샘플)
┌─────────────────────────────────────────────────────────┐
│ 평균: 72.5% | 최대: 89.1% | 최소: 65.2%                  │
│ 경고 임계값: 80.0%                                       │
└─────────────────────────────────────────────────────────┘

📊 트렌드: 📈 증가 추세

⚙️ Configuration

VRAM Optimizer Settings

Parameter	Options	Default	Description
enabled	On/Off	On	Enable/disable automatic VRAM cleaning
clear_mode	Standard/Aggressive	Standard	Cleaning intensity level
auto_clean	Every Time/Only When High	Every Time	Cleaning trigger condition
run_timing	After Queue/Before Queue/Both	After Queue	When to perform cleaning
force_run	0-999	0	Manual trigger (change value to execute)

GPU Monitor Settings

Parameter	Range	Default	Description
monitoring_enabled	On/Off	On	Enable/disable real-time monitoring
update_interval	0.1-10.0	1.0	Update frequency in seconds
history_length	10-300	60	Number of data points to keep
warning_threshold	50.0-95.0	80.0	Memory usage warning percentage
refresh_trigger	0-9999	0	Manual refresh trigger

🔧 Advanced Features

Automatic Dependency Installation

The nodes automatically install required dependencies (GPUtil) on first run.

Background Monitoring

The GPU Monitor runs in a separate thread to provide real-time data without blocking ComfyUI.

ComfyUI Integration

Automatically hooks into ComfyUI's queue execution system for seamless VRAM management.

Terminal Logging

Comprehensive logging system provides detailed information about all operations.

🐛 Troubleshooting

Common Issues

1. "GPU not found" error

   • Ensure GPUtil is installed: pip install GPUtil
   • Check if your GPU is properly detected by your system

2. VRAM cleaning not working

   • Verify PyTorch CUDA support: torch.cuda.is_available()
   • Check if CUDA drivers are properly installed

3. Monitoring not updating

   • Increase the update interval
   • Check ComfyUI terminal for error messages

Getting Help

If you encounter issues:

1. Check the ComfyUI terminal for error messages
2. Open an issue on GitHub with detailed information
3. Join our community discussions

🤝 Contributing

Contributions are welcome! Please feel free to submit a Pull Request.

Development Setup

1. Fork the repository
2. Create a feature branch
3. Make your changes
4. Test thoroughly
5. Submit a pull request

📄 License

This project is licensed under the MIT License - see the LICENSE file for details.

🙏 Acknowledgments

• ComfyUI team for the amazing framework
• GPUtil library for GPU monitoring capabilities
• Community feedback and suggestions

🔗 Links

• ComfyUI
• ComfyUI Manager
• Report Issues

---

Made with ❤️ by StrawberryFist | 🍓 Optimizing your ComfyUI experience!