A fusion computing method and system for cloud platform high-performance computing power and quantum computing power

Abstract

The present invention relates to a method and system for integrating high-performance computing power and quantum computing power of a cloud platform, and belongs to the field of quantum computing technology. The method comprises: submitting a quantum program in a quantum virtual machine through a classical computer interface; performing preliminary classical-quantum task splitting through a compiler to split the task into multiple computing subtasks; inserting a communication primitive before a computing subtask that requires remote invocation, and distributing the corresponding quantum computing subtask to a quantum computing backend and the corresponding classical computing subtask to a high-performance computing power cluster through the communication primitive; executing some classical computing subtasks locally, and simultaneously remotely invoking classical computing subtasks that require the high-performance computing power cluster and quantum computing subtasks that require the quantum computing backend through RoCE communication; and receiving the classical computing subtasks completed by the high-performance computing power cluster and the quantum computing subtasks completed by the quantum computing backend through a resource scheduler.

Description

Technical Field

[0001] The present invention belongs to the field of quantum computing technology, and in particular relates to a method and system for integrating high-performance computing power and quantum computing power of a cloud platform. Background Art

[0002] Quantum computing (QC) has demonstrated quantum superiority over classical computing in tasks such as Shor's integer factorization algorithm, Grover's unstructured database search algorithm, quantum algorithms for linear equations, and quantum adiabatic algorithms. Quantum computers perform computations through quantum acceleration, and their scalability to solve specific problems far exceeds that of even the most powerful classical computers. Key applications of quantum computing include optimization, machine learning, cryptography, and quantum chemistry.

[0003] Quantum computing has reached the Noisy Intermediate-Scale Quantum (NISQ) era, with quantum processing units (QPUs) boasting qubits exceeding 50. With the advent of NISQ hardware, heterogeneous quantum-classical computing architectures have emerged, supporting the efficient execution of a variety of hybrid quantum-classical algorithms.

[0004] Because the construction cost of quantum computers is relatively high, the emergence of the Quantum Computing Cloud Platform has greatly facilitated users' use of quantum computing power.

[0005] However, the scarcity of quantum computing power means users spend a significant amount of time waiting in task queues after submitting computing tasks, resulting in significant inflexibility in task scheduling. Furthermore, due to the bandwidth and latency limitations of traditional cloud network VPCs, users in distributed quantum computing tend to reduce communication and synchronization to improve efficiency, further increasing the inflexibility of task scheduling and limiting the resource sharing and parallel efficiency of quantum cloud architectures. Summary of the Invention

[0006] In view of the above deficiencies in the existing technology, the purpose of the invention is to provide a method and system for the fusion of high-performance computing power and quantum computing power on a cloud platform. The quantum computing backend and the high-performance computing power cluster can be integrated by introducing RoCE communication connections. When processing classical-quantum tasks, quantum computing subtasks and classical computing subtasks are split. When the quantum program is run in a classical computing power environment, some classical computing subtasks are executed locally. At the same time, the classical computing subtasks that require a high-performance computing power cluster and the quantum computing subtasks that require a quantum computing backend are remotely called through RoCE communication. The computing tasks can be reasonably allocated, the queuing time can be reduced, the multi-core high-performance classical computing power can be fully utilized, the scheduling elasticity of classical-quantum hybrid computing tasks can be improved, and the resource sharing capability and parallel efficiency of the cloud architecture can be improved.

[0007] The first aspect of the present invention provides a method for integrating high-performance computing power and quantum computing power on a cloud platform, comprising:

[0008] S1, based on the quantum programming software development kit, submits quantum programs through the classical computer interface in the quantum virtual machine;

[0009] S2, when pre-compiling the quantum program in a classical computing environment, performing preliminary classical-quantum task splitting by the compiler to split the program into multiple computing subtasks, wherein the computing subtasks include quantum computing subtasks and classical computing subtasks;

[0010] S3, inserting a communication primitive before the computing subtask that needs to be remotely called, and distributing the corresponding quantum computing subtask to the quantum computing backend and the corresponding classical computing subtask to the high-performance computing cluster through the communication primitive;

[0011] S4, when running the quantum program in a classical computing environment, executing some classical computing subtasks locally, and remotely calling classical computing subtasks that require the high-performance computing cluster to complete and quantum computing subtasks that require the quantum computing backend to complete through RoCE communication;

[0012] S5, receiving, through a resource scheduler, the classical computing subtasks completed by the high-performance computing cluster and the quantum computing subtasks completed by the quantum computing backend;

[0013] S6, returning the calculation result through the RoCE network, and judging whether the exit condition is met based on the calculation result. If so, exit the classic computing environment; otherwise, return to S4.

[0014] Furthermore, after S2, the fusion computing method of the cloud platform's high-performance computing power and quantum computing power further includes:

[0015] When the quantum program is pre-compiled in a classical computing environment, the multiple split computing subtasks are heuristically scheduled according to the scale of the computing task.

[0016] Furthermore, after S2, the fusion computing method of the cloud platform's high-performance computing power and quantum computing power further includes:

[0017] Generate quantum intermediate representations through the compiler to optimize the instruction execution sequence within each concurrent computing subtask;

[0018] Furthermore, before S4, the fusion computing method of the cloud platform's high-performance computing power and quantum computing power also includes:

[0019] The quantum computer software and hardware interface file of each computing subtask module is generated by the compiler.

[0020] Furthermore, the S5 is specifically as follows:

[0021] The resource scheduler receives the classical computing subtasks completed by the high-performance computing power cluster and the quantum computing subtasks completed by the quantum computing backend at the user's granularity, decides whether to offload some quantum computing subtasks to the high-performance computing power simulator based on the scale of the computing tasks, and schedules the execution order of the quantum computing subtasks on the quantum computing backend.

[0022] The second aspect of the present invention provides a fusion computing system for realizing high-performance computing power and quantum computing power of a cloud platform, which is used to realize the fusion computing method of high-performance computing power and quantum computing power of a cloud platform described in the first aspect, including:

[0023] A submission module, which is used to submit quantum programs through a classical computer interface in a quantum virtual machine based on a quantum programming software development kit;

[0024] a splitting module, configured to perform preliminary classical-quantum task splitting by a compiler when pre-compiling the quantum program in a classical computing environment, thereby splitting the program into a plurality of computing subtasks, wherein the computing subtasks include quantum computing subtasks and classical computing subtasks;

[0025] A distribution module is used to insert a communication primitive before the computing subtask that needs to be remotely called, and distribute the corresponding quantum computing subtask to the quantum computing backend and the corresponding classical computing subtask to the high-performance computing cluster through the communication primitive;

[0026] A calling module is used to execute some classical computing subtasks locally when the quantum program is run in a classical computing environment, and remotely call the classical computing subtasks that require the high-performance computing cluster to complete and the quantum computing subtasks that require the quantum computing backend to complete through RoCE communication;

[0027] A receiving module, configured to receive, through a resource scheduler, the classical computing subtasks completed by the high-performance computing cluster and the quantum computing subtasks completed by the quantum computing backend;

[0028] The exit module is used to return the calculation results through the RoCE network, and determine whether the exit conditions are met based on the calculation results. If so, exit the classic computing environment; otherwise, execute the calling module.

[0029] Furthermore, the cloud platform's integrated computing system combining high-performance computing power and quantum computing power also includes:

[0030] The orchestration module is used to perform heuristic orchestration and scheduling of multiple split computing subtasks according to the scale of the computing task when pre-compiling the quantum program in a classical computing environment.

[0031] Furthermore, the cloud platform's integrated computing system combining high-performance computing power and quantum computing power also includes:

[0032] an optimization module, configured to generate a quantum intermediate representation through the compiler and optimize the instruction execution sequence within each concurrent computing subtask;

[0033] Furthermore, the cloud platform's integrated computing system combining high-performance computing power and quantum computing power also includes:

[0034] A generation module is used to generate a quantum computer hardware and software interface file for each of the computing subtask modules through the compiler.

[0035] Furthermore, the receiving module is specifically configured to:

[0036] The resource scheduler receives the classical computing subtasks completed by the high-performance computing power cluster and the quantum computing subtasks completed by the quantum computing backend at the user's granularity, decides whether to offload some quantum computing subtasks to the high-performance computing power simulator based on the scale of the computing tasks, and schedules the execution order of the quantum computing subtasks on the quantum computing backend.

[0037] The beneficial effects of the present invention are as follows:

[0038] The method and system provided by the present invention integrate the quantum computing backend with the high-performance computing power cluster by introducing the RoCE communication connection. When processing classical-quantum tasks, quantum computing subtasks and classical computing subtasks are split. When the quantum program is run in the classical computing power environment, some classical computing subtasks are executed locally. At the same time, the classical computing subtasks that require the high-performance computing power cluster to complete and the quantum computing subtasks that require the quantum computing backend to complete are remotely called through RoCE communication. This can reasonably allocate computing tasks, reduce queuing time, make full use of multi-core high-performance classical computing power, improve the scheduling flexibility of classical-quantum hybrid computing tasks, and improve the resource sharing capability and parallel efficiency of the cloud architecture. BRIEF DESCRIPTION OF THE DRAWINGS

[0039] The accompanying drawings are only for the purpose of illustrating specific embodiments and are not to be considered as limiting the present invention. Throughout the drawings, the same reference numerals represent the same components. Obviously, the drawings described below are only some of the embodiments of the present invention. Those skilled in the art can also derive other drawings based on these drawings.

[0040] Figure 1 A schematic diagram of a flow chart of a method for integrating high-performance computing power and quantum computing power on a cloud platform provided by an embodiment of the present invention;

[0041] Figure 2 A schematic diagram of the architecture of a method for integrating high-performance computing power and quantum computing power on a cloud platform provided by an embodiment of the present invention;

[0042] Figure 3 A schematic diagram of the structure of a computing system that integrates high-performance computing power and quantum computing power on a cloud platform provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0043] In order to enable those skilled in the art to better understand the technical solutions in the embodiments of the present invention, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the embodiments described are part of the embodiments of the present invention, rather than all of the embodiments. It should be understood that these descriptions are merely exemplary and are not intended to limit the scope of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative work should fall within the scope of protection of the present invention.

[0044] Furthermore, in the following description, descriptions of well-known structures and technologies are omitted to avoid unnecessarily obscuring the concepts disclosed in the present invention.

[0045] In the description of the present invention, it should be noted that, unless otherwise expressly specified and limited, the terms "center", "up", "down", "left", "right", "vertical", "horizontal", "inside", "outside" and the like indicate orientations or positional relationships based on the orientations or positional relationships shown in the accompanying drawings, and are only for the convenience of describing the present invention and simplifying the description, and do not indicate or imply that the system or component referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore cannot be understood as limiting the present invention. In addition, the terms "first", "second" and "third" are used for descriptive purposes only and cannot be understood as indicating or implying relative importance. The terms "installed", "connected" and "connected" should be understood in a broad sense, for example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be a direct connection, or an indirect connection through an intermediate medium, or it can be a communication between the internal parts of two components. For those of ordinary skill in the art, the specific meanings of the above terms in the present invention can be understood according to specific circumstances.

[0046] Exemplary embodiments will be described in detail herein, examples of which are illustrated in the accompanying drawings. In the following description, when referring to the drawings, like numbers in different figures represent the same or similar elements unless otherwise indicated. The embodiments described in the following exemplary embodiments are not intended to represent all possible embodiments consistent with the present invention. Rather, they are merely examples of methods and systems consistent with certain aspects of the present invention, as detailed in the appended claims.

[0047] This paper proposes a method and system for integrating high-performance computing power and quantum computing power into a cloud platform. This approach addresses the scarcity of quantum computing power in existing technologies, which results in users spending significant time queuing in task queues after submitting computing tasks, and the significant rigidity of quantum computing power in task scheduling. Furthermore, due to the bandwidth and latency limitations of traditional cloud network VPCs, users in distributed quantum computing tend to reduce communication and synchronization to improve efficiency, further increasing the rigidity of task scheduling and limiting the resource sharing and parallel efficiency of quantum cloud architectures.

[0048] System Example

[0049] Reference Manual Figure 1 , which shows a flow chart of a method for integrating high-performance computing power and quantum computing power of a cloud platform provided by an embodiment of the present invention.

[0050] Reference Manual Figure 2 , showing an architectural diagram of a method for integrating high-performance computing power and quantum computing power of a cloud platform provided by an embodiment of the present invention.

[0051] The present invention provides a method for integrating high-performance computing power and quantum computing power of a cloud platform, comprising steps S1 to S6.

[0052] S1, based on the quantum programming software development kit, submits quantum programs through the classical computer interface in the quantum virtual machine.

[0053] A software development kit (SDK) is a collection of tools, libraries, and documentation designed to help developers create specific types of software or applications. A programming software development kit (SDK) is a toolkit specifically designed to support the development of programming software (such as programming languages, frameworks, and libraries).

[0054] S2: When pre-compiling a quantum program in a classical computing environment, the compiler performs a preliminary classical-quantum task splitting to produce multiple computing subtasks.

[0055] Among them, computing subtasks include quantum computing subtasks and classical computing subtasks.

[0056] In this invention, splitting computing tasks allows different subtasks to be executed in parallel, thereby fully leveraging multi-core and distributed computing environments and improving overall computing performance. Furthermore, by breaking computing tasks into smaller pieces, available computing resources can be better utilized. Each computing subtask can be executed in parallel on available processors, maximizing computing resource utilization.

[0057] In one possible implementation, after S2, the fusion computing method of the cloud platform's high-performance computing power and quantum computing power further includes:

[0058] When pre-compiling a quantum program in a classical computing environment, the multiple computing subtasks split out are heuristically orchestrated and scheduled based on the scale of the computing task.

[0059] Specifically, each computational subtask can be assigned a scheduling priority based on factors such as task size, complexity, and dependencies. Smaller, simpler, or independent tasks can be given a higher priority. Dependencies between computational subtasks should be analyzed. Some tasks may need to wait until others have completed before they can begin. Scheduling should consider these dependencies to ensure that tasks are executed in the correct order. The potential for parallel execution between tasks should be considered. Some tasks may be able to execute simultaneously, thereby speeding up the overall computational process. Heuristic algorithms can determine the optimal level of parallelism based on resource competition and dependencies between tasks.

[0060] Optionally, the heuristic algorithm may specifically be a greedy algorithm, a genetic algorithm, a particle swarm optimization algorithm, a simulated annealing algorithm, and the like.

[0061] In this invention, heuristic orchestration and scheduling can execute multiple subtasks in parallel based on the scale of the computing task, fully utilizing computing resources and thus improving overall computing performance. Good orchestration and scheduling can ensure that computing resources are effectively utilized, avoiding idle or overloaded resources, and help improve resource utilization efficiency in classic computing environments. By intelligently arranging the execution order of subtasks, the execution time of the entire computing task can be reduced. This is particularly important for tasks that need to be completed quickly, especially in large-scale computing.

[0062] In one possible implementation, after S2, the fusion computing method of the cloud platform's high-performance computing power and quantum computing power further includes:

[0063] The compiler generates quantum intermediate representations and optimizes the instruction execution sequence within each concurrent computing subtask.

[0064] The Quantum Intermediate Representation (QIR) is an abstraction layer used when compiling quantum programs, located between the original quantum program and the hardware instructions of the target quantum computer. QIR provides a unified representation that makes it easier for compilers to optimize, analyze, and convert, while adapting to different quantum computer architectures.

[0065] Specifically, the compiler translates the quantum program into an intermediate representation, the quantum intermediate representation. For the instruction sequence within each concurrent computational subtask, the compiler can perform instruction-level optimizations. Specifically, it can perform local optimizations within a basic block (a set of linearly arranged instructions), such as constant propagation and dead code removal. It also optimizes loop structures, such as loop unrolling and loop variable substitution. It analyzes the dependencies between instructions to determine which instructions can be executed in parallel, thereby improving concurrency. Furthermore, for quantum computing, there is concurrent control flow. The compiler can perform some optimizations, specifically rearranging instructions to fully utilize computing resources and improve pipeline efficiency. It ensures that, whenever possible, different quantum gate instructions can be executed concurrently at the same time.

[0066] In the present invention, by optimizing the instruction execution sequence within each concurrent computational subtask, more efficient concurrent execution can be achieved, thereby improving overall computing performance. Simultaneously, by optimizing the instruction sequence, the execution time of each computational subtask can be reduced, helping to reduce the execution time of the overall computational task, particularly in the case of large-scale concurrent computations. Furthermore, optimizing the instruction sequence can reduce the communication overhead between computational subtasks. In quantum computing, some instruction sequences may involve communication between qubits, and reasonable optimization can reduce the cost of this communication.

[0067] S3 inserts a communication primitive before the computing subtask that needs to be called remotely, distributes the corresponding quantum computing subtask to the quantum computing backend through the communication primitive, and distributes the corresponding classical computing subtask to the high-performance computing power cluster.

[0068] High-Performance Computing (HPC) is a computer technology that uses powerful computing capabilities to solve complex and large-scale problems. HPC systems are designed to provide higher processing speeds and computing power than conventional computer systems to handle scientific, engineering, and commercial applications that require large amounts of computing resources and storage capacity.

[0069] Communication primitives refer to the basic operations or instructions used to support communication and data transmission in the computing model.

[0070] Specifically, the communication original words may refer to the Verbs API.

[0071] In this invention, by distributing different types of computing tasks to specially designed backends, we can make better use of their respective resources. The quantum computing backend is used to perform quantum computing tasks, while the high-performance computing cluster is used for classical computing tasks, making optimal use of each computing resource.

[0072] S4, when running a quantum program in a classical computing environment, executes some classical computing subtasks locally, and at the same time remotely calls classical computing subtasks that require a high-performance computing cluster and quantum computing subtasks that require a quantum computing backend through RoCE communication.

[0073] RoCE (RDMA over Converged Ethernet) communication is a technology that implements RDMA (Remote Direct Memory Access) on Ethernet networks. RDMA allows direct memory access between two computers on a network without the intervention of the central processing unit (CPU). RoCE is a way to implement RDMA on Ethernet and is commonly used to provide low-latency and high-bandwidth data transmission.

[0074] In this invention, by executing some classical computing subtasks locally while remotely invoking high-performance computing clusters and quantum computing backends to complete other tasks, different computing resources can be synergistically utilized to achieve optimal resource allocation. At the same time, leveraging the computing power of high-performance computing clusters and quantum computing backends can significantly improve the computing performance of the entire system. This distributed computing approach helps accelerate the completion of computing tasks. Furthermore, executing some classical computing subtasks locally while remotely invoking other computing resources can reduce the waiting time for the overall computing task. While tasks are being executed locally, other parts can be executed in parallel on remote clusters or quantum computing backends, improving efficiency.

[0075] In one possible implementation, before S4, the fusion computing method of the cloud platform's high-performance computing power and quantum computing power further includes:

[0076] The quantum computer hardware and software interface files for each computing subtask module are generated by the compiler.

[0077] In this invention, generating software and hardware interface files helps abstract and standardize the hardware details of the underlying quantum computer. Furthermore, generating standard interface files makes it easier to run quantum programs on different quantum computers, improving cross-platform compatibility. The compiled interface files are adaptable to quantum computers of different types and sizes, allowing the same quantum program to be executed on computers with different architectures.

[0078] S5 receives the classical computing subtasks completed by the high-performance computing cluster and the quantum computing subtasks completed by the quantum computing backend through the resource scheduler.

[0079] In a possible implementation, S5 specifically includes: receiving the classical computing subtasks completed by the high-performance computing power cluster and the quantum computing subtasks completed by the quantum computing backend at the user's granularity through the resource scheduler, deciding whether to offload some quantum computing subtasks to the high-performance computing power simulator based on the scale of the computing tasks, and scheduling the execution order of the quantum computing subtasks on the quantum computing backend.

[0080] In the present invention, the resource scheduler receives the classical computing subtasks completed by the high-performance computing power cluster and the quantum computing subtasks completed by the quantum computing backend at the user's granularity. The scheduler receives and processes tasks at the user's granularity, which can more effectively coordinate the use of the resources of the high-performance computing power cluster and the quantum computing backend, minimize resource waste, and improve the resource utilization efficiency of the entire system.

[0081] Furthermore, by deciding whether to offload to a high-performance computing simulator based on task size, we can optimize computing task performance and ensure optimal resource matching. Offloading some tasks to a high-performance computing simulator can reduce the burden on the quantum computing backend and alleviate its computational pressure. This is beneficial for ensuring backend stability and reliability. Furthermore, different computing tasks may have different requirements for computing resources. Dynamically adjusting the execution location and method of tasks helps adapt to different computing needs and improves the flexibility of the entire system.

[0082] S6: The calculation results are returned via the RoCE network. Based on the calculation results, it is determined whether the exit conditions are met. If so, the system exits the classic computing environment. Otherwise, the system returns to S4.

[0083] The exit condition may specifically be in accordance with classic control logic.

[0084] In this invention, by integrating HPC classical computing power with a scalable RDMA storage cluster, the quantum circuit classical simulator can flexibly offload small- and medium-scale quantum computing tasks. Furthermore, due to the hybrid scheduling of HPC and quantum computing power, quantum-classical hybrid algorithms can be efficiently executed. Therefore, compared to traditional quantum cloud platform implementations, this invention can enhance the scale-of-service capabilities of quantum cloud platforms and improve the scheduling flexibility and execution efficiency of quantum-classical hybrid computing tasks.

[0085] The beneficial effects of the present invention are as follows:

[0086] The method and system provided by the present invention integrate the quantum computing backend with the high-performance computing power cluster by introducing the RoCE communication connection. When processing classical-quantum tasks, quantum computing subtasks and classical computing subtasks are split. When the quantum program is run in the classical computing power environment, some classical computing subtasks are executed locally. At the same time, the classical computing subtasks that require the high-performance computing power cluster to complete and the quantum computing subtasks that require the quantum computing backend to complete are remotely called through RoCE communication. This can reasonably allocate computing tasks, reduce queuing time, make full use of multi-core high-performance classical computing power, improve the scheduling flexibility of classical-quantum hybrid computing tasks, and improve the resource sharing capability and parallel efficiency of the cloud architecture.

[0087] System Example

[0088] Reference Manual Figure 3 , which shows a structural diagram of a fusion computing system of high-performance computing power and quantum computing power of a cloud platform provided by an embodiment of the present invention.

[0089] Another specific embodiment of the present invention discloses a fusion computing system 20 for cloud platform high-performance computing power and quantum computing power, which is used to implement the fusion computing method of cloud platform high-performance computing power and quantum computing power in the system embodiment, including:

[0090] A submission module 201 is used to submit a quantum program in a quantum virtual machine through a classical computer interface based on a quantum programming software development kit;

[0091] a splitting module 202 for performing preliminary classical-quantum task splitting by a compiler when pre-compiling the quantum program in a classical computing environment, and splitting the task into a plurality of computing subtasks, wherein the computing subtasks include quantum computing subtasks and classical computing subtasks;

[0092] The distribution module 203 is used to insert a communication primitive before the computing subtask that needs to be remotely called, and distribute the corresponding quantum computing subtask to the quantum computing backend and the corresponding classical computing subtask to the high-performance computing cluster through the communication primitive;

[0093] A calling module 204 is configured to execute some classical computing subtasks locally when the quantum program is run in a classical computing environment, and remotely call classical computing subtasks that require the high-performance computing cluster to complete and quantum computing subtasks that require the quantum computing backend to complete via RoCE communication.

[0094] A receiving module 205 is configured to receive, through a resource scheduler, the classical computing subtasks completed by the high-performance computing cluster and the quantum computing subtasks completed by the quantum computing backend;

[0095] The exit module 206 is used to return the calculation result through the RoCE network, and determine whether the exit condition is met based on the calculation result. If so, exit the classic computing environment; otherwise, execute the calling module.

[0096] In one possible implementation, the fusion computing system of the cloud platform's high-performance computing power and quantum computing power further includes:

[0097] The orchestration module is used to perform heuristic orchestration and scheduling of multiple split computing subtasks according to the scale of the computing task when pre-compiling the quantum program in a classical computing environment.

[0098] In one possible implementation, the fusion computing system of the cloud platform's high-performance computing power and quantum computing power further includes:

[0099] an optimization module, configured to generate a quantum intermediate representation through the compiler and optimize the instruction execution sequence within each concurrent computing subtask;

[0100] In one possible implementation, the fusion computing system of the cloud platform's high-performance computing power and quantum computing power further includes:

[0101] A generation module is used to generate a quantum computer hardware and software interface file for each of the computing subtask modules through the compiler.

[0102] In a possible implementation, the receiving module 205 is specifically configured to:

[0103] The resource scheduler receives the classical computing subtasks completed by the high-performance computing power cluster and the quantum computing subtasks completed by the quantum computing backend at the user's granularity, decides whether to offload some quantum computing subtasks to the high-performance computing power simulator based on the scale of the computing tasks, and schedules the execution order of the quantum computing subtasks on the quantum computing backend.

[0104] The fusion computing system 20 of cloud platform high-performance computing power and quantum computing power provided by the present invention can execute the above-mentioned fusion computing system of cloud platform high-performance computing power and quantum computing power, and achieve the same or similar technical effects. To avoid repetition, the present invention will not go into details.

[0105] The beneficial effects of the present invention are as follows:

[0106] The method and system provided by the present invention integrate the quantum computing backend with the high-performance computing power cluster by introducing the RoCE communication connection. When processing classical-quantum tasks, quantum computing subtasks and classical computing subtasks are split. When the quantum program is run in the classical computing power environment, some classical computing subtasks are executed locally. At the same time, the classical computing subtasks that require the high-performance computing power cluster to complete and the quantum computing subtasks that require the quantum computing backend to complete are remotely called through RoCE communication. This can reasonably allocate computing tasks, reduce queuing time, make full use of multi-core high-performance classical computing power, improve the scheduling flexibility of classical-quantum hybrid computing tasks, and improve the resource sharing capability and parallel efficiency of the cloud architecture.

[0107] Finally, it should be noted that the above embodiments are merely illustrative of the technical solutions of the embodiments of the present invention, and are not intended to limit them. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they may still modify the technical solutions described in the aforementioned embodiments, or replace some of the technical features therein with equivalents; and such modifications or replacements do not deviate from the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention. Any changes or replacements that can be easily conceived by those skilled in the art within the technical scope disclosed in the present invention should be covered by the scope of protection of the present invention.

Claims

1. A method for integrating high-performance computing power and quantum computing power of a cloud platform, characterized in that: include: S1, based on the quantum programming software development kit, submits quantum programs through the classical computer interface in the quantum virtual machine; S2, when pre-compiling the quantum program in a classical computing environment, performing preliminary classical-quantum task splitting by the compiler to split the program into multiple computing subtasks, wherein the computing subtasks include quantum computing subtasks and classical computing subtasks; S3, inserting a communication primitive before the computing subtask that needs to be remotely called, and distributing the corresponding quantum computing subtask to the quantum computing backend and the corresponding classical computing subtask to the high-performance computing cluster through the communication primitive; S4, when running the quantum program in a classical computing environment, executing some classical computing subtasks locally, and remotely calling classical computing subtasks that require the high-performance computing cluster to complete and quantum computing subtasks that require the quantum computing backend to complete through RoCE communication; S5, receiving, through a resource scheduler, the classical computing subtasks completed by the high-performance computing cluster and the quantum computing subtasks completed by the quantum computing backend; S6, returning the calculation result through the RoCE network, and judging whether the exit condition is met based on the calculation result. If so, exit the classic computing environment; otherwise, return to S4.

2. The method for integrating high-performance computing power and quantum computing power of a cloud platform according to claim 1 is characterized in that: After S2, the method further includes: When the quantum program is pre-compiled in a classical computing environment, the multiple split computing subtasks are heuristically scheduled according to the scale of the computing task.

3. The fusion computing method of cloud platform high-performance computing power and quantum computing power according to claim 1 is characterized in that: After S2, the method further includes: The compiler generates a quantum intermediate representation and optimizes the instruction execution sequence within each concurrent computing subtask.

4. The method for integrating high-performance computing power and quantum computing power of a cloud platform according to claim 1, characterized in that: Before the S4, it also includes: The quantum computer software and hardware interface file of each computing subtask module is generated by the compiler.

5. The method for integrating high-performance computing power and quantum computing power of a cloud platform according to claim 1, characterized in that: The S5 is specifically: The resource scheduler receives the classical computing subtasks completed by the high-performance computing power cluster and the quantum computing subtasks completed by the quantum computing backend at the user's granularity, decides whether to offload some quantum computing subtasks to the high-performance computing power simulator based on the scale of the computing tasks, and schedules the execution order of the quantum computing subtasks on the quantum computing backend.

6. A fusion computing system of high-performance computing power and quantum computing power of cloud platform, characterized by: A method for realizing the fusion of high-performance computing power and quantum computing power of a cloud platform as claimed in any one of claims 1 to 5, comprising: A submission module, which is used to submit quantum programs through a classical computer interface in a quantum virtual machine based on a quantum programming software development kit; a splitting module, configured to perform preliminary classical-quantum task splitting by a compiler when pre-compiling the quantum program in a classical computing environment, thereby splitting the program into a plurality of computing subtasks, wherein the computing subtasks include quantum computing subtasks and classical computing subtasks; A distribution module is used to insert a communication primitive before the computing subtask that needs to be remotely called, and distribute the corresponding quantum computing subtask to the quantum computing backend and the corresponding classical computing subtask to the high-performance computing cluster through the communication primitive; A calling module is used to execute some classical computing subtasks locally when the quantum program is run in a classical computing environment, and remotely call the classical computing subtasks that require the high-performance computing cluster to complete and the quantum computing subtasks that require the quantum computing backend to complete through RoCE communication; A receiving module, configured to receive, through a resource scheduler, the classical computing subtasks completed by the high-performance computing cluster and the quantum computing subtasks completed by the quantum computing backend; The exit module is used to return the calculation results through the RoCE network, and determine whether the exit conditions are met based on the calculation results. If so, exit the classic computing environment; otherwise, execute the calling module.

7. The fusion computing system of cloud platform high-performance computing power and quantum computing power according to claim 6 is characterized in that: Also includes: The orchestration module is used to perform heuristic orchestration and scheduling of multiple split computing subtasks according to the scale of the computing task when pre-compiling the quantum program in a classical computing environment.

8. The fusion computing system of cloud platform high-performance computing power and quantum computing power according to claim 6 is characterized in that: Also includes: An optimization module is used to generate a quantum intermediate representation through the compiler and optimize the instruction execution sequence in each concurrent computing subtask.

9. The fusion computing system of cloud platform high-performance computing power and quantum computing power according to claim 6 is characterized in that: Also includes: A generation module is used to generate a quantum computer hardware and software interface file for each of the computing subtask modules through the compiler.

10. The fusion computing system of cloud platform high-performance computing power and quantum computing power according to claim 6, characterized in that: The receiving module is specifically used for: The resource scheduler receives the classical computing subtasks completed by the high-performance computing power cluster and the quantum computing subtasks completed by the quantum computing backend at the user's granularity, decides whether to offload some quantum computing subtasks to the high-performance computing power simulator based on the scale of the computing tasks, and schedules the execution order of the quantum computing subtasks on the quantum computing backend.

Images