A cloud scheduling system and method across quantum technology routes

Abstract

This invention discloses a cloud scheduling system and method across quantum technology routes, relating to the field of quantum computing cloud platform technology. The system uses a unified hardware abstraction module to perform quintuple-based standardized abstraction of heterogeneous quantum processing units such as superconducting, ion traps, and photonic quantum units. Combined with a constraint resolution and scheduling decision module, it completes multi-level matching and scheduling of tasks and hardware based on physical constraints such as the number of qubits, connectivity, gate fidelity, and coherence time. Finally, a hardware adaptation execution module realizes instruction conversion and task execution. This invention can shield the underlying differences between different quantum hardware, achieving unified scheduling and efficient utilization of quantum resources from multiple technology routes, improving the compatibility and task execution success rate of the quantum computing cloud platform.

Description

Technical Field

[0001] This invention relates to the field of quantum computing cloud platform technology, specifically to a hardware abstraction and task scheduling method and system compatible with different quantum processing units such as superconducting, ion traps, and optical quantum. Background Technology

[0002] Most existing quantum computing cloud platforms are designed for only a single type of quantum processing unit. The instruction sets, physical constraints, topologies, and qubit characteristics of different quantum hardware vary significantly, making it impossible to flexibly schedule and adapt user tasks across different types of quantum hardware. Furthermore, traditional cloud computing scheduling strategies do not consider the unique physical limitations of quantum tasks, such as coherence time constraints, qubit connectivity constraints, and quantum gate fidelity, making it difficult to achieve efficient matching of quantum tasks with heterogeneous quantum hardware. The lack of a unified hardware abstraction layer and a quantum cloud scheduling mechanism that adapts to multiple technical approaches further limits the hardware compatibility and resource utilization of quantum computing cloud platforms. Summary of the Invention

[0003] To address the issues of poor hardware compatibility and inability to schedule quantum tasks across different technology routes in existing quantum computing cloud platforms, this invention provides a cloud scheduling system and method that spans quantum technology routes. By unifying the hardware abstraction model and constraint matching scheduling rules, it achieves unified management of different types of quantum processing units and adaptive scheduling of quantum computing tasks.

[0004] A cloud scheduling system across quantum technology routes includes: a quantum task access module, a unified hardware abstraction module, a constraint resolution module, a scheduling decision module, a hardware adaptation execution module, and a heterogeneous quantum hardware cluster; the heterogeneous quantum hardware cluster includes at least two of the following: superconducting quantum processing units, ion trap quantum processing units, and optical quantum processing units.

[0005] The Unified Hardware Abstraction Module employs a quintuple abstraction model to standardize the description of heterogeneous quantum hardware. The quintuple is defined as: QHA = {QN, QC, QF, QT, QL}, where QN is the number of qubits, QC is the qubit connectivity constraint, QF is the quantum gate fidelity, QT is the quantum coherence time, and QL is the supported instruction set type. The Unified Hardware Abstraction Module performs real-time information acquisition and standardized quintuple encapsulation of heterogeneous quantum processing units, shielding the underlying implementation differences between different quantum hardware and providing a unified hardware view.

[0006] The constraint resolution module is used to resolve the constraints of user-submitted quantum tasks, including: minimum number of qubits required, target topology, minimum quantum gate fidelity threshold, maximum acceptable execution delay, and specified quantum technology route preferences.

[0007] The scheduling decision module executes a constraint-matching scheduling strategy. First, based on the minimum number of qubits required by the quantum task, it selects candidate quantum hardware that satisfies QN ≥ the required number of qubits. Then, based on the task topology requirements, it matches the QC connectivity constraints of the candidate hardware, retaining the set of topology-compatible hardware. Subsequently, based on the fidelity threshold specified by the task, it selects quantum hardware with QF not lower than the threshold. Combining QT coherence time and the estimated execution time of the task, it eliminates hardware that fails to execute the task due to insufficient coherence time. If multiple hardware meet the conditions, they are sorted from high to low quantum gate fidelity, and the optimal hardware is selected. If the user specifies a technical route preference, the available hardware of the corresponding route is matched first.

[0008] Based on the scheduling decision, the hardware adaptation execution module converts the unified format quantum task instructions into the native instruction set of the target quantum hardware, sends them to the corresponding quantum processing unit for execution, and sends the execution results back to the user terminal.

[0009] A cloud scheduling method across quantum technology routes includes the following steps: S1. A quantum task access module receives user quantum computing tasks and extracts task constraint parameters; S2. A unified hardware abstraction module collects heterogeneous quantum hardware states and generates standardized quintuple hardware abstraction information; S3. A constraint resolution module performs structured parsing of task constraints and hardware abstraction information; S4. A scheduling decision module performs multi-level constraint matching to determine the target quantum processing unit; S5. A hardware adaptation execution module completes instruction conversion and task distribution and execution; S6. After execution, the computation results are collected and fed back to the user.

[0010] This invention achieves unified abstraction and management of quantum hardware using multiple technological approaches, including superconductivity, ion traps, and photonic quantum computing. It enhances the hardware compatibility of quantum cloud platforms, performs precise scheduling based on the unique physical constraints and task requirements of quantum hardware, improves the success rate and resource utilization of quantum tasks, shields the differences in underlying hardware, and reduces the development and adaptation costs for users of heterogeneous quantum computing resources. Detailed Implementation

[0011] Taking a cloud platform containing superconducting quantum processing units and ion trap quantum processing units as an example: the unified hardware abstraction module collects two types of hardware information and encapsulates them into quintuples; the user submits a quantum task requiring 10 qubits, a linear topology, and a dual-quantum-gate fidelity ≥99%; after the constraint resolution module extracts the constraints, the scheduling decision module selects ion trap quantum hardware that meets the requirements for quantity, topology, fidelity, and coherence time; the hardware adaptation execution module converts the general quantum task instructions into the native instructions of the hardware and issues them for execution, finally returning the calculation results.

Claims

1. A cloud scheduling system that crosses quantum technology routes, characterized in that, include: The system comprises a quantum task access module, a unified hardware abstraction module, a constraint parsing module, a scheduling decision module, a hardware adaptation execution module, and a heterogeneous quantum hardware cluster; the heterogeneous quantum hardware cluster includes at least two of the following: a superconducting quantum processing unit, an ion trap quantum processing unit, and a photonic quantum processing unit. The unified hardware abstraction module is used to standardize the encapsulation of heterogeneous quantum hardware using a five-tuple abstraction model. The five-tuple is QHA={QN, QC, QF, QT, QL}, where QN is the number of qubits, QC is the qubit connectivity constraint, QF is the quantum gate fidelity, QT is the quantum coherence time, and QL is the supported instruction set type. The scheduling decision module is used to perform multi-level constraint matching based on the quantum task constraints and the abstract information of the quintuple, and select the target quantum processing unit from the heterogeneous quantum hardware cluster.

2. The system according to claim 1, characterized in that, The constraint resolution module is used to resolve the constraints of quantum tasks, including minimum number of qubits required, target topology, minimum quantum gate fidelity threshold, maximum acceptable execution delay, and quantum technology route preference.

3. The system according to claim 1, characterized in that, The matching steps performed by the scheduling decision module include: Based on the minimum number of bits required for quantum tasks, candidate quantum hardware that satisfies QN ≥ the required number of bits is selected. Based on the task topology requirements, match the QC connectivity constraints of candidate hardware and retain the set of topology-compatible hardware. Based on the fidelity threshold specified in the task, select quantum hardware with a QF not lower than the threshold. By combining QT coherence time with the estimated execution time of the task, hardware with insufficient coherence time is eliminated; The remaining hardware is sorted from highest to lowest quantum gate fidelity to select the optimal hardware, or the corresponding quantum hardware is matched preferentially according to the user's technical route preference.

4. The system according to claim 1, characterized in that, The hardware adaptation execution module is used to convert the unified format quantum task instructions into the native instruction set of the target quantum processing unit, and to issue execution and return results.

5. A cloud scheduling method across quantum technology routes, characterized in that, Applied to the system of any one of claims 1 to 4, comprising the following steps: S1. The quantum task access module receives user quantum computing tasks and extracts the corresponding task constraint parameters; S2. The unified hardware abstraction module collects the state information of each quantum processing unit in the heterogeneous quantum hardware cluster and generates standardized quintuple hardware abstraction information; S3. The constraint parsing module performs structured parsing of task constraint parameters and quintuple hardware abstraction information; S4. The scheduling decision module determines the target quantum processing unit that is suitable for the current quantum task based on multi-level constraint matching rules; S5. The hardware adaptation execution module converts the quantum task into the native instructions corresponding to the target quantum processing unit and issues them for execution; S6. Receive the execution results of the quantum hardware and send them back to the user terminal.