Ray Distributed Matrix Calculation System

This project demonstrates a distributed matrix calculation system using Ray framework, where matrix operations are distributed across multiple services.

System Architecture

• client_a.py: Main client program that:

  • Initializes Ray cluster
  • Deploys B and C services
  • Generates initial matrix
  • Coordinates the distributed calculation
  • Collects final result

• serve_b.py: B Service that:

  • Performs matrix multiplication with its fixed matrix
  • Forwards result to C Service

• serve_c.py: C Service that:

  • Performs final matrix multiplication
  • Returns the result to client

Requirements

• Python 3.7+
• Ray 2.0+
• NumPy

Installation

1. Install Python dependencies:

pip install ray numpy

2. Clone this repository (if applicable)

Running the System

1. Start the system by running:

python3 client_a.py

2. Expected output flow:

• Ray cluster initialization
• Services deployment
• Matrix calculation process
• Final result display

Expected Output

You should see logs showing:

1. Ray cluster initialization
2. Services deployment
3. Matrix operations at each step
4. Final calculated matrix result

Notes

• The system uses Ray's default ports (10001 for client, 8265 for dashboard)
• Dashboard is available at: http://127.0.0.1:8265
• For production use, modify the Ray initialization to connect to a proper cluster instead of local mode