English | 中文
over-engineering memory holder
- PID controller
- Asymmetric strategy (fast release, conservative allocation)
- Scene-aware scoring (5 scenarios)
- Adaptive parameter optimization
- Performance tracking
- Benchmark system (9 test scenarios including nonlinear patterns)
pip install -r requirements.txt# Run all tests
python -m pytest tests/ -v
# Run specific tests
python -m pytest tests/test_controllers.py -v# Standard mode (dynamic target, random 25-35%)
python run_holder.py
# Fixed target mode
python run_holder.py --fixed-target 80
# Disable benchmark export
python run_holder.py --no-benchmarkpython run_benchmark.pyQuick Install (recommended):
curl -fsSL https://raw.githubusercontent.com/bOOOOcG/nerdy-holder/main/remote-install.sh | sudo bashAuto Install (skip prompts, use defaults):
AUTO=yes curl -fsSL https://raw.githubusercontent.com/bOOOOcG/nerdy-holder/main/remote-install.sh | sudo bashManual Install:
git clone https://github.com/bOOOOcG/nerdy-holder.git
cd nerdy-holder
sudo bash install.shUninstall:
curl -fsSL https://raw.githubusercontent.com/bOOOOcG/nerdy-holder/main/remote-uninstall.sh | sudo bashAuto Uninstall (skip confirmation):
CONFIRM=yes curl -fsSL https://raw.githubusercontent.com/bOOOOcG/nerdy-holder/main/remote-uninstall.sh | sudo bashMonitoring:
systemctl status nerdy-holder # Service status
bash deployment/monitor.sh # Monitoring dashboardChange target memory usage: Edit /etc/systemd/system/nerdy-holder.service, modify --fixed-target parameter, then:
sudo systemctl daemon-reload
sudo systemctl restart nerdy-holderNote: Must run directly on host machine. Docker containers not supported (isolated memory namespace). Requirements: Linux (Ubuntu 20.04+), Python 3.8+, root privileges.
nerdy-holder/
├── nerdy_holder/ # Core package
│ ├── controllers/ # PID and response calculators
│ ├── predictors/ # EMA predictors
│ ├── optimizers/ # Parameter optimizers
│ ├── trackers/ # Performance trackers
│ ├── memory/ # Memory block management
│ └── core.py # Core program
├── tests/ # Test modules
│ ├── benchmark/ # Benchmark system
│ └── test_*.py # Unit tests (35 tests)
├── deployment/ # Deployment scripts
├── run_holder.py # Holder entry point
└── run_benchmark.py # Benchmark entry point
- Kp: Proportional control
- Ki: Integral control (with asymmetric recovery)
- Kd: Derivative control
- Release: Fast response, low cost
- Allocation: Conservative filling, high cost
- 35% Error
- 30% Stability
- 20% Block rate
- 15% Adjustment rhythm