FRACTIS

Logo

FRACTIS

Home

GitHub

FRACTIS TESTNET

FRACTIS TESTNET is the test network for our revolutionary decentralized database system. It provides a safe environment for testing and development of decentralized database applications.

Core Features

Decentralized Architecture

Fully distributed system with no single point of failure.

Smart Contract Integration

Seamless integration with smart contracts for automated data handling.

High Performance

Optimized for speed and efficiency in data operations.

Security First

Built with robust security measures at its core.

Scalable Design

Horizontal scaling with dynamic node allocation and adaptive load balancing.

Cross-Chain Compatibility

Seamless interoperability with multiple blockchain networks and protocols.

Node Requirements

Hardware Requirements


  • CPU: 2+ cores

  • RAM: 4GB minimum

  • Storage: 50GB SSD minimum

  • Network: Stable internet connection with minimum 5Mbps up/down

Software Requirements


  • Operating System: Ubuntu 20.04+ / macOS 12+ / Windows 10/11

  • Rust 1.70.0 or higher

  • Git

Lightweight LLM Support

FRACTIS TESTNET optionally supports running lightweight LLM capabilities. This is an advanced feature that requires additional hardware resources.

Additional Hardware Requirements for LLM


  • CPU: 6+ cores recommended

  • RAM: 16GB minimum

  • Storage: Additional 10GB SSD

  • GPU: NVIDIA GPU with 6GB+ VRAM (optional, for better performance)

  • Additional Software: CUDA Toolkit 11.8+ (if using GPU)

LLM Specifications


  • Model: LLaMA-2 7B v2.0

  • Version: 2.0.1 (December 2023)

  • Quantization: 4-bit GPTQ quantization

  • Memory Usage: ~8GB RAM when active

  • Disk Space: ~5GB for model files

  • Response Time: 2-3 seconds for short responses

  • Context Window: 4096 tokens

  • Language Support: Multilingual (40+ languages)

Node Setup Guide

Install Rust and Clone Repository

Copy

# Install Rust

curl --proto '=https' --tlsv1.2 -sSf https://sh.rustup.rs | sh

# Clone the repository

git clone https://github.com/FractisAI/FRACTIS-TESTNET

cd fractis-node

Node Configuration (config/node.toml)

Copy

# Basic node configuration

node_id = "auto" # Will be automatically generated

host = "0.0.0.0" # Listen on all interfaces

port = 8000

storage_path = "./data"

max_connections = 50

consensus_timeout = 5000 # Milliseconds

bootstrap_nodes = [

"testnet.fractis.io:8000",

"testnet2.fractis.io:8000"

]

# Optional LLM configuration (disabled by default)

[llm]

enabled = false # Set to true to enable LLM features

model_path = "./models/llama-2-7b-q4.safetensors"

tokenizer_path = "./models/tokenizer.json"

max_batch_size = 4

use_gpu = false # Set to true if using GPU

Start Your Node

Copy

# For basic node (without LLM):

# Build and start basic node

cargo build --release

./target/release/fractis-node --config config/node.toml

# For node with LLM support (optional):

# First, download LLM model files (about 5GB)

./scripts/download_models.sh

# Build with GPU support

cargo build --release --features llm,cuda

# Or build with CPU-only LLM support

cargo build --release --features llm

# Start node with LLM enabled (modify config.toml first)

./target/release/fractis-node --config config/node.toml

Performance Optimization

Basic Node Optimization


  • Monitor disk space usage

  • Adjust network bandwidth usage

  • Configure max_connections based on available resources

LLM Optimization


  • Start with CPU-only mode first

  • Enable GPU support if available

  • Adjust batch size based on available memory

  • Consider running during off-peak hours

Troubleshooting

Basic Node Issues


  • Check network connectivity

  • Verify bootstrap nodes are accessible

  • Ensure port 8000 is open

LLM Issues (if enabled)


  • Insufficient memory: Disable LLM feature or upgrade RAM

  • Slow response: Consider enabling GPU support

  • High CPU usage: Reduce batch size or disable LLM

Diagnostic Command

Copy

# Run node diagnostics

./fractis-node diagnose

# This command will check system compatibility

# and report potential configuration issues

Support

Technical Support


  • Check the Documentation

  • Open an issue on GitHub