Overall Architecture Diagram
---------------------------

The overall architecture of Origin PilotOS is shown in the following figure.

.. figure:: image/ScreenShot_2026-03-18_133855_964.png
    :align: center
    :width: 100%

**Application Layer**: Mainly refers to the user programming space, where users can develop various applications on top of the Origin PilotOS system.

**Application Development Toolkits**: Origin PilotOS provides multiple development toolkits for upper-level users, which are already integrated into the Pilot system. After deploying the Origin PilotOS system, users can directly conduct secondary function development or quickly verify quantum algorithms based on these toolkits. The currently included toolkits are:
ChemiQ (quantum computing chemistry software package),
QFinancial (quantum finance development toolkit),
QCFD (quantum computing fluid dynamics development toolkit),
Quantum Heuristic Algorithm Toolkit, and
System Manager Interface (system management API).

**Built-in Programming Frameworks**: The Origin PilotOS system provides three integrated programming frameworks:


a. Quantum Programming Framework.
A framework dedicated to quantum program development, providing: 

1) High-level quantum programming language; 
2) Simulation interface for executing and verifying quantum algorithms; Advanced quantum circuit tools, including circuit printing, equivalence checking, circuit reversal, and conversion to other quantum programming languages (such as Qiskit, Cirq).

b. Quantum-Classical Hybrid Compilation Framework.
Optimizes hybrid quantum-classical programs through: 

1) Analyzing and scheduling the execution order between quantum tasks and classical tasks; 
2) Applying compiler-level optimizations to improve overall algorithm efficiency and resource utilization.

c. Quantum Machine Learning (QML) Framework.
Supports seamless development of quantum and hybrid machine learning models, featuring: 

1) Python-based user-friendly API with support for automatic differentiation and dynamic computation graphs; 
2) Unified execution across multiple hardware backends (quantum processors, CPU, GPU); 
3) Built-in high-performance numerical libraries for accelerating training and inference; 
4) Compatibility with mainstream machine learning frameworks (such as PyTorch, TensorFlow), facilitating integration into existing ML workflows.

**System Interface**: The unified entry point for upper-level applications to access the Origin PilotOS system. It enforces isolation between the system layer and application layer, ensuring that user space applications can only access system resources in a limited and controlled manner, thereby maintaining system stability and security.

**Computing Service Layer**: Handles specific business scenarios with dedicated processing logic, decoupling application-level computing tasks from core system computing to ensure overall computing stability. Currently integrated services include:

1) Quantum State Tomography (QST), 
2) Expectation Value Estimation, 
3) Error Mitigation (EM) Computing, 
4) Distributed Quantum Circuit Computing.

**System Service Layer**: The Origin PilotOS quantum operating system is built on a set of core, interoperable modules: system management module, scheduling service module, compilation service module, resource management module, and backend service module. These modules communicate seamlessly through secure internal system channels to implement coordinated workflow execution—from task submission and compilation to resource allocation and quantum computing on physical hardware. This architecture design with separation of concerns and loose coupling ensures high reliability, scalability, and compatibility across heterogeneous quantum backends.

**Computing Backends**: Origin PilotOS natively supports multiple quantum computing backends, including superconducting quantum processors, quantum simulators, and classical computing backends.