uhcr 5.1.0

Universal Hardware-Aware Compute Runtime — Enterprise-grade modular execution stack with C++ safety checking

UHCR - Universal Hardware-Aware Compute Runtime

Enterprise-Grade Modular Execution Stack with C++ Safety Checking

UHCR is a production-ready Python framework for hardware-optimized computation with comprehensive safety checking, JIT compilation, and enterprise deployment tools.

🚀 Key Features

🛡️ Enterprise Safety (NEW in v5.0)

• C++ Safety Checkers - Native memory bounds, overflow detection, division-by-zero prevention
• Runtime Protection - Stack overflow, heap exhaustion, execution timeout enforcement
• 17 Violation Types - Comprehensive dangerous operation detection
• Python Fallbacks - Full safety even without native library

⚡ Performance

• JIT Compilation - Traces Python functions and compiles to native machine code
• Multi-Backend - CPU (AVX512/AVX2/SSE), CUDA, Metal, ROCm
• SIMD Optimization - Automatic vectorization and hardware feature detection
• Near-Zero Overhead - ~1-2ns safety check latency with native library

🔧 Complete CLI Toolkit (18 Commands)

• uhcr compile - AOT compilation with integrity checking
• uhcr docker/k8s - Generate deployment manifests
• uhcr hw - Comprehensive hardware detection
• uhcr safety - Runtime safety management
• uhcr mcp_start - AI agent integration (MCP server)
• ...and more!

🏢 Enterprise Ready

• Provenance Tracking - Full build metadata and SHA-256 checksums
• Container Support - Docker and Kubernetes manifest generation
• Plugin System - Extensible architecture with TOML manifests
• Type Safety - Full type hints with py.typed marker

📦 Installation

# Basic installation
pip install uhcr

# With enterprise features
pip install uhcr[enterprise]

# With all features
pip install uhcr[all]

# Build native safety library (optional but recommended)
uhcr build

🎯 Quick Start

Basic JIT Compilation

import uhcr

@uhcr.jit(eager=True)
def compute(a, b):
    return (a + b) * 2

result = compute(5, 3)  # Compiled on first call

With Safety Checking

from uhcr.security import enable_safety_checks, safe_add, safe_mul

# Enable runtime safety
enable_safety_checks()

# Safe operations - automatically checked
result = safe_add(a, b)  # Raises SafetyViolation on overflow
result = safe_mul(x, y)  # Protected multiplication

Compile for Production

# Compile Python to native with integrity checks
uhcr compile myapp.py --optimize 3

# Creates myapp.uhcrc/ with:
#   • source.py - Original code
#   • metadata.json - Build provenance
#   • checksums.json - SHA-256 integrity
#   • __main__.py - Executable runner
#   • README.md - Documentation

# Run compiled module
python myapp.uhcrc/

Hardware Detection

# Full hardware report
uhcr hw

# Just the fingerprint
uhcr hw --fingerprint
# Output: Windows-AMD64-avx2-cuda_12.4+vulkan

Container Deployment

# Generate Dockerfile
uhcr docker myapp.py --image mycompany/app:v1

# Generate Kubernetes manifest
uhcr k8s myapp.py --image mycompany/app:v1 \
    --replicas 5 \
    --cpu-request 500m \
    --memory-request 1Gi

AI Agent Integration

# Start MCP server for AI agents
uhcr mcp_start --transport http --port 3000

# AI agents can now use UHCR tools:
#   • compile_code
#   • optimize_code
#   • detect_hardware
#   • run_benchmark
#   • generate_docker/k8s

📋 CLI Commands

uhcr -v                    # Version
uhcr hw                    # Hardware detection
uhcr compile script.py     # Compile to native
uhcr docker script.py      # Generate Dockerfile
uhcr k8s script.py         # Generate K8s manifest
uhcr safety status         # Safety system status
uhcr safety test           # Test safety features
uhcr build                 # Build C++ safety library
uhcr run script.py --jit   # Run with JIT
uhcr optimize script.py    # Optimize code
uhcr mcp_start             # Start MCP server
# ... 18 total commands

🏗️ Architecture

Core Components

• Compiler - IR-based multi-backend code generation
• Runtime - JIT execution and memory management
• Hardware - CPUID, GPU detection, NUMA topology
• Safety - C++ bounds checking, overflow detection
• Backends - CPU (AVX512/AVX2), CUDA, Metal, ROCm
• Storage - Memory pooling and hierarchical caching
• Network - gRPC/HTTP server, distributed coordination

Safety System

uhcr/native/
├── safety_checker.hpp     # C++ safety header
├── safety_checker.cpp     # C++ implementation
├── CMakeLists.txt         # Build config
└── build_native.py        # Python build script

uhcr/security/
├── runtime_safety.py      # Python bindings
└── __init__.py            # Public API

🛡️ Safety Features

Memory Safety

• ✅ Buffer overflow/underflow prevention
• ✅ Null pointer dereference detection
• ✅ Array bounds validation
• ✅ Use-after-free detection

Arithmetic Safety

• ✅ Integer overflow/underflow detection
• ✅ Division by zero prevention
• ✅ Safe casting validation

Resource Safety

• ✅ Stack overflow protection (10K recursion limit)
• ✅ Heap exhaustion prevention (1GB limit)
• ✅ Execution timeout (30s default)
• ✅ Infinite loop detection

Complete Safety Documentation →

📊 Performance

# Benchmark example
uhcr benchmark --suite tensor --output results.json

# Expected performance:
# - Native safety checks: ~1-2ns overhead
# - Python fallbacks: ~10-50ns overhead
# - JIT compilation: 10-100x speedup vs Python

🏢 Enterprise Use Cases

Financial Services

• High-frequency trading with safety guarantees
• Risk calculation with overflow protection
• Regulatory compliance (provenance tracking)

Healthcare

• Medical imaging processing
• Patient data analysis with safety checks
• HIPAA-compliant deployments

AI/ML

• Model training with GPU acceleration
• Inference optimization
• AI agent integration via MCP

DevOps

• Automated code optimization
• Container generation
• CI/CD integration

📚 Documentation

• Full Documentation
• Safety Guide
• CLI Reference
• API Reference
• Changelog

🤝 Contributing

We welcome contributions! See CONTRIBUTING.md for guidelines.

# Development setup
git clone https://github.com/VishveshJoshi89/UHCR
cd UHCR
pip install -e .[dev]
uhcr build
pytest tests/

📄 License

Apache License 2.0

🔗 Links

• PyPI: https://pypi.org/project/uhcr/
• Repository: https://github.com/VishveshJoshi89/UHCR
• Documentation: https://vishveshjoshi89.github.io/UHCR-DOCS/
• Bug Tracker: https://github.com/VishveshJoshi89/UHCR/issues
• Changelog: https://github.com/VishveshJoshi89/UHCR/blob/main/CHANGELOG.md

🌟 Star History

If you find UHCR useful, please consider starring the repository!

---

Built with ❤️ for enterprise Python developers