Raven Python Miner

High-Performance Distributed KAWPOW (Ravencoin) GPU Miner with Stratum Coordinator, HTTP DAG Streaming, and Multi-GPU Support

---

📖 Overview

Raven Python Miner is an asynchronous, distributed mining engine built specifically for the Ravencoin (RVN) KAWPOW proof-of-work algorithm.

Designed for both single-machine miners and large-scale cloud GPU fleets (such as Vast.ai, RunPod, and Lambda Labs), this miner decouples mining pool coordination from GPU hash execution:

                  ┌───────────────────────────────┐
                  │    Stratum Pool (2Miners)     │
                  └───────────────┬───────────────┘
                                  │ Stratum v1 TCP
                                  ▼
                  ┌───────────────────────────────┐
                  │   KAWPOW Server Coordinator   │
                  │   - Pools connections         │
                  │   - Synthesizes / Caches DAG  │
                  │   - HTTP DAG Streaming (:8080)│
                  │   - TLS Job Dispatcher (:8088)│
                  └───────────────┬───────────────┘
                                  │ TLS 1.3 / TCP
       ┌──────────────────────────┼──────────────────────────┐
       ▼                          ▼                          ▼
┌──────────────┐           ┌──────────────┐           ┌──────────────┐
│ GPU Worker 1 │           │ GPU Worker 2 │           │ GPU Worker N │
│ (RTX 3090)   │           │ (RTX 3060)   │           │ (RTX 4090)   │
└──────────────┘           └──────────────┘           └──────────────┘

---

✨ Key Features

• Consensus-Identical KAWPOW Engine: Implements the official Ravencoin ProgPOW specification, including Keccak-f[800], Keccak-f[1600], KISS99 PRNG, dynamic 64-round unrolled execution, and memory-hard DAG access.
• Distributed Fleet Architecture: Run one lightweight server coordinator that manages pool jobs, share validation, and dispatching, while dozens of remote GPU instances execute searches.
• Centralized HTTP DAG Streaming: The server synthesizes the multi-gigabyte DAG (5.73 GB at Epoch 605) once and serves it to client workers via high-speed HTTP streaming. Workers avoid redundant generation and start hashing within seconds.
• Native TLS 1.3 Encryption: Secure worker-to-server communications over public cloud IPs with automatic self-signed X.509 certificate generation.
• Dynamic Epoch Transitions: Automatically detects block height epoch rollovers, generates the new DAG, updates worker VRAM, and resumes mining without downtime or manual restarts.
• Full Multi-GPU & Nonce Partitioning: Spreads nonce search spaces automatically across all designated GPUs per worker rig.
• Real-Time NVML Telemetry: Live reporting of per-GPU hashrate, compute utilization, memory utilization, temperature, and power consumption.

---

⚡ Hashrate Performance Reference

Typical performance on calibrated KAWPOW settings:

GPU Model	Architecture	VRAM	Hashrate	Power	Efficiency
NVIDIA GeForce RTX 3060	Ampere	12 GB GDDR6	~21 – 23 MH/s	~125 W	~0.18 MH/W
NVIDIA Quadro RTX 6000	Turing	24 GB GDDR6	~38 – 42 MH/s	~220 W	~0.19 MH/W
NVIDIA GeForce RTX 3080	Ampere	10 GB GDDR6X	~47 – 50 MH/s	~250 W	~0.20 MH/W
NVIDIA GeForce RTX 3090	Ampere	24 GB GDDR6X	~58 – 62 MH/s	~300 W	~0.20 MH/W
NVIDIA GeForce RTX 4090	Ada Lovelace	24 GB GDDR6X	~75 – 82 MH/s	~320 W	~0.24 MH/W

---

🚀 Quick Start

For detailed installation instructions on Linux, Windows, and Cloud GPU instances, see INSTALL.md.

1. Installation

# Clone the repository
git clone https://github.com/APS-PPR-2016/raven_python_miner.git
cd raven_python_miner

# Install required dependencies
pip install -r requirements.txt

# for a quickstart
pip install numpy pycuda cuda-python cryptography

2. Launching Single-Machine Mining (Local Server + GPU Worker)

Step 1: Start Coordinator in Terminal 1

python3 distributed/kawpow_server.py \
  --pool-host rvn.2miners.com \
  --pool-port 6060 \
  --wallet RNGyc94iUJnYWCcshAobk2xjUYw7EvLsq4 \
  --rig-name worker_1 \
  --port 8088 \
  --http-port 8080 \
  --no-tls

Step 2: Start GPU Miner in Terminal 2

python3 distributed/kawpow_client.py \
  --server-host 127.0.0.1 \
  --server-port 8088 \
  --http-url http://127.0.0.1:8080 \
  --gpus 0 \
  --worker-name worker_1_gpu0 \
  --no-tls

---

☁️ Distributed Cloud Mining (Vast.ai / RunPod / AWS)

When renting cloud GPUs with high bandwidth:

1. Host Coordinator on your master server or VPS:

   python3 distributed/kawpow_server.py \
     --pool-host rvn.2miners.com \
     --pool-port 6060 \
     --wallet YOUR_RVN_WALLET_ADDRESS \
     --rig-name cloud_rig \
     --port 8088 \
     --http-port 8080

   (By default, TLS 1.3 is enabled with auto-generated certificates in server.crt).

2. Run GPU Clients on any number of rented cloud nodes:

   python3 distributed/kawpow_client.py \
     --server-host <COORDINATOR_IP> \
     --server-port 8088 \
     --http-url http://<COORDINATOR_IP>:8080 \
     --gpus 0,1 \
     --worker-name vast_node_1

---

⚙️ CLI Reference

Coordinator Server (kawpow_server.py)

Parameter	Default	Description
--pool-host	rvn.2miners.com	Stratum mining pool hostname
--pool-port	6060	Stratum pool port
--wallet	Required	Ravencoin payout address
--rig-name	default_rig	Rig identifier sent to pool
--port	8088	Client coordinator TCP/TLS port
--http-port	8080	DAG binary HTTP download port
--no-tls	False	Disables TLS encryption (plaintext TCP)
--cache-dir	Current Dir	Directory to store generated DAG cache files

Client Worker (kawpow_client.py)

Parameter	Default	Description
--server-host	localhost	Coordinator server IP or hostname
--server-port	8088	Coordinator server port
--http-url	http://localhost:8080	Coordinator HTTP DAG download URL
--gpus	0	Comma-separated GPU indices (e.g. 0,1,2)
--batch-size	524288	Number of nonces searched per GPU batch
--worker-name	worker_1	Worker name reported in logs and telemetry
--no-tls	False	Connects over plaintext TCP without TLS

---

📂 Repository Structure

raven_python_miner/
├── distributed/               # Distributed Python Mining System
│   ├── engine_core.py         # KAWPOW math, JIT ProgPOW compiler, Keccak DAG kernels
│   ├── kawpow_server.py       # Stratum pool listener, HTTP DAG server, share dispatcher
│   ├── kawpow_client.py       # Multi-GPU worker, NVML telemetry, VRAM DAG loader
│   └── tls_utils.py           # Native TLS 1.3 context & self-signed certificate generation
├── tests/                     # Test suite & verification scripts
│   ├── test_context.py        # CUDA context initialization check
│   ├── test_nvrtc.py          # NVRTC JIT compilation test
│   ├── test_dag_search_benchmark.py # Batch benchmark
│   └── test_stratum_duplex.py # Stratum protocol duplex communication test
├── docs/                      # Architectural and technical documentation
├── INSTALL.md                 # Complete installation and deployment guide
├── requirements.txt           # Python package dependencies
├── pyproject.toml             # Project build configuration
└── README.md                  # Project overview & documentation

---

🔒 Security

• Worker-to-server communications support TLS 1.2 / TLS 1.3.
• Automatically generates 2048-bit RSA keys and X.509 certificates with Subject Alternative Names (localhost, standard IP addresses).
• Custom Certificate Authorities (--tls-ca) are supported for enterprise deployments.

---

📄 License

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

---

⚠️ Disclaimer

This software is intended for educational, testing, and research purposes in cryptocurrency mining and GPU compute architectures. Ensure compliance with your local laws, power regulations, and hosting provider terms of service when running mining workloads.

✨ Donate

Generous donations are welcome to RNGyc94iUJnYWCcshAobk2xjUYw7EvLsq4 or a coffee keeps me awake to do more..

https://buymeacoffee.com/aps2016