Getting Started with RHAPSODY¶
Welcome to RHAPSODY! This section will guide you through everything you need to know to start using RHAPSODY for your heterogeneous HPC-AI workflows.
What is RHAPSODY?¶
RHAPSODY is a high-performance runtime system designed for executing complex workflows that combine traditional HPC simulations with AI/ML workloads. It provides a unified interface for orchestrating tasks across different computational paradigms and computing resources.
Prerequisites¶
Before getting started with RHAPSODY, ensure you have:
- Python 3.9 or higher
- Basic familiarity with async/await programming
- Understanding of workflow concepts
- Access to computational resources (local machine or HPC cluster)
Learning Path¶
Follow this recommended learning path to master RHAPSODY:
1. Installation¶
Start by installing RHAPSODY and its dependencies.
2. Quick Start¶
Learn the basics with a simple "Hello World" workflow.
3. First Workflow¶
Build your first real workflow with multiple tasks.
4. Configuration¶
Learn how to configure RHAPSODY for your specific needs.
Key Concepts¶
Before diving into the tutorials, familiarize yourself with these key concepts:
Sessions¶
A session represents a workflow execution context that manages the lifecycle of tasks and resources.
Backends¶
Backends are pluggable execution engines that determine how and where your tasks run: - Dask: For local and distributed Python computing - RADICAL-Pilot: For large-scale HPC execution on supercomputers and clusters
Tasks¶
Tasks are the fundamental units of computation in RHAPSODY workflows, defined with executables, arguments, and dependencies.
Callbacks¶
Callbacks provide real-time monitoring and response to task state changes during workflow execution.
Getting Help¶
If you encounter issues or have questions:
- API Reference: Consult the detailed API Reference
- Issues: Report bugs or ask questions on GitHub Issues
- Community: Join our community discussions on GitHub
Next Steps¶
Ready to begin? Start with the Installation Guide to set up RHAPSODY on your system.
Once you're comfortable with the basics, explore the API Reference for detailed documentation of all available functions and classes.