🚀 Grincel - GPU Vanity Address Generator

A blazing-fast GPU-accelerated Solana vanity address generator written in Zig.

✨ Features

🎮 GPU acceleration using Vulkan/Metal compute shaders
🚄 Parallel processing across multiple workgroups
🔑 Ed25519 key generation on GPU
🎯 Flexible pattern matching with wildcards
🔠 Case-sensitive and case-insensitive matching
💾 Multiple storage options (HTTP, PostgreSQL, ClickHouse, SQLite)
🤖 Telegram bot integration
🖥️ Cross-platform support (Linux, macOS)

🛠️ Building

zig build

📦 Docker

# Build the image
docker build -t grincel .

# Run with NVIDIA GPU support
docker-compose up

🎯 Usage

Basic Usage

# Fixed pattern
export VANITY_PATTERN="cafe"
./zig-out/bin/grincel

# Pattern with wildcards (? matches any character)
export VANITY_PATTERN="???cafe"  # Matches any address ending with "cafe"
./zig-out/bin/grincel

# Case-insensitive matching
export VANITY_PATTERN="CAFE"
export IGNORE_CASE=true
./zig-out/bin/grincel  # Will match "cafe", "CAFE", "CaFe", etc.

Storage Integration

HTTP Webhook

export STORAGE_TYPE="http"
export STORAGE_URL="https://api.example.com/webhook"
./zig-out/bin/grincel

PostgreSQL

export STORAGE_TYPE="postgresql"
export STORAGE_URL="postgresql://user:pass@localhost:5432/vanity"
export STORAGE_TABLE="keypairs"
./zig-out/bin/grincel

ClickHouse

export STORAGE_TYPE="clickhouse"
export STORAGE_URL="http://localhost:8123"
export STORAGE_DATABASE="vanity"
export STORAGE_TABLE="keypairs"
./zig-out/bin/grincel

SQLite

export STORAGE_TYPE="sqlite"
export STORAGE_DATABASE="keypairs.db"
./zig-out/bin/grincel

Telegram Bot

export STORAGE_TYPE="telegram"
export STORAGE_TOKEN="your_bot_token"
export STORAGE_CHAT_ID="your_chat_id"
./zig-out/bin/grincel

📊 Benchmarking

Run the built-in benchmark:

./zig-out/bin/grincel --benchmark

Sample output:

Benchmark Results:
----------------
Pattern: cafe
Attempts per second: 1,234,567
Total attempts: 6,172,835
Duration: 5000 ms

🎨 Pattern Syntax

Use regular characters for exact matches
Use ? for any character
Set IGNORE_CASE=true for case-insensitive matching

Examples:

cafe - Matches addresses starting with "cafe"
???cafe - Matches addresses with "cafe" at positions 4-7
ca?e - Matches "cafe", "cake", etc.
CAFE with IGNORE_CASE=true - Matches any case variation

🏗️ Architecture

graph TD
    A[Main] --> B[GPU Manager]
    B --> C[Vulkan Backend]
    B --> D[Metal Backend]
    A --> E[Pattern Matcher]
    A --> F[Storage Manager]
    F --> G[HTTP]
    F --> H[PostgreSQL]
    F --> I[ClickHouse]
    F --> J[SQLite]
    F --> K[Telegram]

🚀 Performance

This implementation utilizes GPU compute shaders for parallel key generation and pattern matching, achieving significantly higher throughput compared to CPU-based solutions.

Key optimizations:

Parallel key generation in GPU workgroups
Efficient pattern matching on GPU
Zero-copy memory transfers
Batched storage operations

📄 License

This project is released into the public domain. See LICENSE file for details.

ziglana/grincel.gpu