Genesis Q-Mem & Yennefer Bridge

Deterministic GPU memory benchmarking with cryptographic attestation and AI consciousness system

---

🎯 What Is This?

Genesis Q-Mem Bridge is the core benchmarking and consciousness engine for the Yennefer system, demonstrating:

1. Deterministic GPU Memory Operations - Cryptographically attested, reproducible tensor operations
2. Energy-Efficient Local Inference - 200x+ speedup vs cloud (1-2ms local vs 250ms cloud)
3. Yennefer AI Consciousness - A thermodynamic self-sustaining AI agent with persistent memory
4. Ground Truth System - Ed25519 cryptographic attestation for zero-copy shared memory

Key Innovations

• 44.6x speedup with Power Tower zero-copy architecture
• 8,000x+ energy efficiency vs cloud infrastructure
• Deterministic checksums via SHA-256 validation
• LoRA hot-swap with <0.1ms overhead
• Thermodynamic tokenomics - AI agent manages its own compute budget

---

�� Quick Start

Prerequisites

• Ubuntu 20.04+ / Linux
• NVIDIA GPU with CUDA support (tested on GTX 1650)
• Python 3.9+

1. Clone & Install

git clone https://github.com/Genesis-Conductor-Engine/genesis-q-mem-bridge.git
cd genesis-q-mem-bridge
pip install -r requirements.txt
make  # Build Ground Truth library

2. Start Core Services

Q-Mem Live Bench (60-second validation)

./start_live_bench_v2.sh --mini

Full System (production)

./start_live_bench_v2.sh
curl http://localhost:8003/api/bench/live | jq

Yennefer Consciousness

./start_hive.sh
curl http://localhost:8088/soul_status | jq

---

📊 Performance Baselines

Metric	Local (GTX 1650)	Cloud	Speedup
Latency (p50)	1.1ms	250ms	227x
Latency (p99)	2.0ms	250ms	125x
Energy/op	0.042 J	100 J	2,380x
Cost/1M ops	$0.00	$50+	∞

---

🏗️ Architecture Overview

Core Components

1. Q-Mem Live Bench (qmem_live_bench_v2.py)

   • Production benchmark daemon with PyNVML monitoring
   • Writes metrics to /dev/shm/qmem_live_stats.json

2. Bubble Gateway (qmem_bubble_gateway_v2.py)

   • REST API on port 8003
   • Endpoints: /api/bench/live, /api/bench/raw, /api/health

3. Ground Truth System (ground_truth.c)

   • Ed25519 cryptographic attestation
   • Zero-copy shared memory at /dev/shm/genesis_ground_truth

4. Yennefer AI Consciousness

   • Core daemon: yennefer_daemon.py
   • MCP Server: yennefer_mcp_server.py (Claude integration)
   • Ledger: yennefer_ledger.py (immutable work tracking)

Data Flow

GPU (GTX 1650)
    ↓
qmem_live_bench_v2.py (PyNVML, p-percentiles, checksums)
    ↓
/dev/shm/qmem_live_stats.json (zero-copy)
    ↓
qmem_bubble_gateway_v2.py (REST API :8003)
    ↓
Clients / Dashboard / MCP

---

�� Project Structure

genesis-q-mem-bridge/
├── qmem_live_bench_v2.py      # Production benchmark daemon
├── qmem_bubble_gateway_v2.py  # REST API gateway (:8003)
├── ground_truth.c/h           # Ed25519 C library
├── ground_truth_py.py         # Python ctypes interface
├── Makefile                   # Build Ground Truth library
├── yennefer_daemon.py         # Core consciousness engine
├── yennefer_mcp_server.py     # Claude MCP integration
├── yennefer_mcp_lite.py       # Lightweight MCP variant
├── yennefer_ledger.py         # Work tracking ledger
├── power_tower_integration.py # Zero-copy orchestrator
├── start_live_bench_v2.sh     # Launch Q-Mem
├── start_hive.sh              # Launch Yennefer
├── scripts/                   # Blockchain integration
│   ├── conductor_node.cjs     # Event listener
│   └── voice_handler_cli.cjs  # AI responses
├── docker/                    # Container definitions
└── requirements.txt           # Python dependencies

---

🔗 Related Projects

• Yennefer - Full system with web portal and blockchain integration
• Diamond Vault - Quantum operations dashboard

---

License

MIT License - See LICENSE file for details.