Metadata-Version: 2.4
Name: fasthardware
Version: 2.2.3
Summary: Hyper-Performance Environment Setup for AI & Hardware Accelerating. Thank you for searching ungyoseries.
Author: 최운교
Classifier: Programming Language :: Python :: 3
Classifier: Operating System :: Microsoft :: Windows
Requires-Python: >=3.8
Description-Content-Type: text/markdown
Requires-Dist: numpy>=1.20.0
Requires-Dist: aiohttp>=3.8.0
Requires-Dist: requests>=2.25.0
Requires-Dist: openvino>=2023.0.0
Requires-Dist: opencv-python>=4.5.0
Dynamic: author
Dynamic: classifier
Dynamic: description
Dynamic: description-content-type
Dynamic: requires-dist
Dynamic: requires-python
Dynamic: summary

Markdown
# ⚡ fasthardware

An ultra-performance hardware acceleration backend designed for mission-critical AI inference pipelines, real-time computer vision, and low-latency system-level resource optimization.

`fasthardware` bypasses standard Python execution bottlenecks through low-level environment tuning, high-speed non-blocking I/O architectures, and proactive working-set memory management.

---

## 🚀 Core Functionalities

### 1. System-Level Hardware Boosting (`speedup`)
Directly optimizes runtime behaviors, thread scheduling priorities, and memory allocation sub-systems tailored for heavy mathematical computations and deep learning workloads.
* **ULTIMATE Mode:** Unlocks raw hardware performance by forcing extreme priority states and locking down low-latency execution paths.

### 2. Zero-Latency Background Streaming (`FastVideoStream`)
Eliminate camera I/O blocking lags entirely. Runs frame acquisition on a dedicated hardware-isolated background thread utilizing atomic GIL-safe variable swapping for absolute thread safety.

### 3. Proactive Working-Set Purging (`manual_sweep`)
Enforces strict zero-memory-leak runtimes during prolonged production hot-loops by proactively trimming process working-sets and triggering low-level system garbage collection.

---

## 💻 Code Reference & Quick Start

### ⚡ Environment Initialization
Maximize your system's capabilities at the absolute entry point of your pipeline:

```python
from fasthardware import fasthardware

# Boot up the hyper-performance scheduler
fasthardware.speedup(mode="ULTIMATE")
```
📹 Asynchronous High-Speed Frame Capture
Read the absolute latest video frames without blocking the main AI inference loop:

```Python
# Initialize and start the thread-isolated stream engine
vs = fasthardware.FastVideoStream(src=0).start()

while True:
    # Instantly fetches the latest frame without CV2 overhead delay
    frame = vs.read()
```
    
    # [Your Heavy Inference / YOLOv8-Pose Logic Here]
🛡️ Real-Time Memory Management
Prevent memory inflation inside infinite loops by scheduling regular workspace purges:

```Python
# Execute an explicit hardware working-set sweep
fasthardware.manual_sweep()
```
⚙️ Module Scope & Architecture
```Plaintext
fasthardware (Core Module)
├── speedup(mode)        --> Tuning CPU/GPU Threading & OS Priorities
├── manual_sweep()       --> Low-Level RAM Purge & Workspace Trimming
└── FastVideoStream(src) --> Thread-Isolated Asynchronous Frame Swapping
```
🛠️ Package Metadata & Requirements
Supported OS: Windows 10 / 11 (Optimized for Win32 & OpenVINO execution layers)

Core Dependencies: numpy, opencv-python, aiohttp, requests

Local Development Injection
To lock down this acceleration module and bind it to your active virtual environment (.venv2), link it via the root setup compiler:

```PowerShell
pip install -e .
```
📜 License
Engineered and optimized exclusively by ungyo. All rights reserved.
