WO2022105440A1 - 一种量子与经典混合云平台以及任务执行方法 - Google Patents

一种量子与经典混合云平台以及任务执行方法 Download PDF

Info

Publication number
WO2022105440A1
WO2022105440A1 PCT/CN2021/121221 CN2021121221W WO2022105440A1 WO 2022105440 A1 WO2022105440 A1 WO 2022105440A1 CN 2021121221 W CN2021121221 W CN 2021121221W WO 2022105440 A1 WO2022105440 A1 WO 2022105440A1
Authority
WO
WIPO (PCT)
Prior art keywords
quantum
classical
task
computing
layer
Prior art date
Application number
PCT/CN2021/121221
Other languages
English (en)
French (fr)
Inventor
李红珍
张新
赵雅倩
李仁刚
Original Assignee
苏州浪潮智能科技有限公司
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by 苏州浪潮智能科技有限公司 filed Critical 苏州浪潮智能科技有限公司
Priority to US18/020,616 priority Critical patent/US20230297401A1/en
Publication of WO2022105440A1 publication Critical patent/WO2022105440A1/zh

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F9/00—Arrangements for program control, e.g. control units
    • G06F9/06—Arrangements for program control, e.g. control units using stored programs, i.e. using an internal store of processing equipment to receive or retain programs
    • G06F9/44—Arrangements for executing specific programs
    • G06F9/455—Emulation; Interpretation; Software simulation, e.g. virtualisation or emulation of application or operating system execution engines
    • G06F9/45533—Hypervisors; Virtual machine monitors
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N10/00—Quantum computing, i.e. information processing based on quantum-mechanical phenomena
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F8/00—Arrangements for software engineering
    • G06F8/40—Transformation of program code
    • G06F8/41—Compilation
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F9/00—Arrangements for program control, e.g. control units
    • G06F9/06—Arrangements for program control, e.g. control units using stored programs, i.e. using an internal store of processing equipment to receive or retain programs
    • G06F9/46—Multiprogramming arrangements
    • G06F9/54—Interprogram communication
    • G06F9/545—Interprogram communication where tasks reside in different layers, e.g. user- and kernel-space
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N10/00—Quantum computing, i.e. information processing based on quantum-mechanical phenomena
    • G06N10/60—Quantum algorithms, e.g. based on quantum optimisation, quantum Fourier or Hadamard transforms
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N10/00—Quantum computing, i.e. information processing based on quantum-mechanical phenomena
    • G06N10/80—Quantum programming, e.g. interfaces, languages or software-development kits for creating or handling programs capable of running on quantum computers; Platforms for simulating or accessing quantum computers, e.g. cloud-based quantum computing
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N20/00—Machine learning
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/04—Architecture, e.g. interconnection topology
    • G06N3/0464—Convolutional networks [CNN, ConvNet]

Definitions

  • the present application relates to the field of cloud computing technology, in particular to a quantum and classical hybrid cloud platform and a task execution method.
  • Quantum computing is one of the most promising avenues to solve the above problems.
  • the quantum cloud platform will become the main form of existence of quantum computing in the future.
  • the quantum computing cloud platform mainly provides online quantum chips or simulation services.
  • classical computing needs to be completed locally or in other settings, and then frequent communication with the quantum computing cloud platform is carried out.
  • Communication to complete the entire calculation due to the need for frequent communication between classical computing clusters and quantum computing clusters, cross-cluster communication causes a lot of communication overhead, resulting in excessive data delay between quantum chips and classical devices, and even loss of original quantum advantage.
  • the purpose of this application is to provide a quantum and classical hybrid cloud platform and a task execution method, which can reduce communication overhead and data delay, improve task processing efficiency, and take advantage of quantum computing. Its specific plan is as follows:
  • the present application discloses a quantum and classical hybrid cloud platform, including:
  • the SaaS layer is used to provide a user interface, so as to obtain the quantum and classical hybrid programming language corresponding to the task to be executed through the user interface;
  • the PaaS layer is used to perform algorithm compilation and task separation on the quantum and classical hybrid programming language, obtain quantum computing tasks and classical computing tasks corresponding to the tasks to be executed, and allocate the quantum computing tasks and the classical computing tasks respectively. resource;
  • the IaaS layer is configured to use a quantum virtual machine to perform the quantum computing task and use a classical server to perform the classical computing task according to the resource allocation situation of the PaaS layer.
  • the SaaS layer includes:
  • the user programming module is used to provide a user interface, so as to obtain the quantum and classical hybrid programming language corresponding to the task to be executed through the user interface.
  • the resource management and scheduling module is used for:
  • the PaaS layer includes:
  • the quantum virtual machine deployment module is configured to acquire the information of the first target classical server determined by the resource management and scheduling module, and deploy the quantum virtual machine on the first target classical server.
  • the IaaS layer includes:
  • the second target classical server is used for executing the classical computing task.
  • the IaaS layer includes:
  • a network device used for communication between different devices in the IaaS layer.
  • the IaaS layer includes:
  • the infrastructure management module is used to manage, monitor and operate the infrastructure of the IaaS layer.
  • the SaaS layer includes:
  • Solutions provide modules for delivering machine vision solutions and reinforcement learning solutions.
  • the present application discloses a quantum and classical hybrid task execution method, which is applied to the aforementioned quantum and classical hybrid cloud platform, including:
  • the quantum virtual machine is used to perform the quantum computing task and the classical server is used to perform the classical computing task through the IaaS layer according to the resource allocation situation of the PaaS layer.
  • this application discloses a quantum and classical hybrid cloud platform
  • the cloud platform includes a SaaS layer for providing a user interface, so as to obtain the quantum and classical hybrid programming language corresponding to the task to be executed through the user interface;
  • the PaaS layer For performing algorithm compilation and task separation on the quantum and classical hybrid programming language, obtaining quantum computing tasks and classical computing tasks corresponding to the tasks to be performed, and assigning resources to the quantum computing tasks and the classical computing tasks respectively;
  • the IaaS layer is configured to use a quantum virtual machine to perform the quantum computing task and use a classical server to perform the classical computing task according to the resource allocation situation of the PaaS layer.
  • the present application uses a user interface at the SaaS layer, so that users can input a hybrid quantum and classical programming language through the user interface, which solves the user-unfriendly problem caused by a single mode that only supports quantum programming in the existing quantum cloud platform , and when the PaaS layer compiles the quantum and classical hybrid programming languages, the tasks to be executed are divided into quantum computing tasks and classical computing tasks, and the corresponding IaaS layer resources are configured to execute the corresponding tasks, realizing the dual computing mode. Simultaneous and fast execution, maximum utilization of computing resources, and improved task processing efficiency.
  • both the quantum virtual machine for quantum computing and the classical virtual machine for classical computing are located at the IaaS layer, making quantum virtual machines for quantum computing and
  • the communication between classical virtual machines for classical computing becomes intra-cluster communication, which reduces the delay of cross-cluster communication, reduces communication overhead and data delay, and exerts the advantages of quantum computing.
  • 1 is a schematic structural diagram of a quantum and classical hybrid cloud platform disclosed in the application.
  • FIG. 2 is a schematic structural diagram of a specific quantum and classical hybrid cloud platform disclosed in the application.
  • FIG. 3 is a schematic structural diagram of a specific quantum and classical hybrid cloud platform disclosed in the application.
  • FIG. 4 is a flowchart of a specific quantum and classical hybrid task execution method disclosed in the present application.
  • the quantum computing cloud platform mainly provides online quantum chips or simulation services.
  • classical computing needs to be completed locally or in other settings, and then frequent communication with the quantum computing cloud platform is carried out.
  • the communication completes the entire calculation.
  • the cross-cluster communication causes a lot of communication overhead, which causes the data delay between the quantum chip and the classical device to be too large, and even loses the original data.
  • Quantum advantage proposes a quantum and classical hybrid cloud platform, which can reduce communication overhead and data delay, improve task processing efficiency, and take advantage of quantum computing.
  • an embodiment of the present application discloses a quantum-classical hybrid cloud platform, which includes:
  • the SaaS layer 11 is used to provide a user interface, so as to obtain the quantum and classical hybrid programming language corresponding to the task to be executed through the user interface;
  • the PaaS layer 12 is used to perform algorithm compilation and task separation on the quantum and classical hybrid programming language, and obtain quantum computing tasks and classical computing tasks corresponding to the tasks to be executed, which are respectively the quantum computing tasks and the classical computing tasks resource allocation;
  • the IaaS layer 13 is configured to use a quantum virtual machine to perform the quantum computing task and use a classical server to perform the classical computing task according to the resource allocation situation of the PaaS layer.
  • this application discloses a quantum and classical hybrid cloud platform
  • the cloud platform includes a SaaS layer for providing a user interface, so as to obtain a quantum and classical hybrid programming language corresponding to a task to be executed through the user interface;
  • the PaaS layer For performing algorithm compilation and task separation on the quantum and classical hybrid programming language, obtaining quantum computing tasks and classical computing tasks corresponding to the tasks to be performed, and assigning resources to the quantum computing tasks and the classical computing tasks respectively;
  • the IaaS layer is configured to use a quantum virtual machine to perform the quantum computing task and use a classical server to perform the classical computing task according to the resource allocation situation of the PaaS layer.
  • the present application uses a user interface at the SaaS layer, so that users can input a hybrid quantum and classical programming language through the user interface, which solves the user-unfriendly problem caused by a single mode that only supports quantum programming in the existing quantum cloud platform , and when the PaaS layer compiles the quantum and classical hybrid programming languages, the tasks to be executed are divided into quantum computing tasks and classical computing tasks, and the corresponding IaaS layer resources are configured to execute the corresponding tasks, realizing the dual computing mode. Simultaneous and fast execution, maximum utilization of computing resources, and improved task processing efficiency.
  • both the quantum virtual machine for quantum computing and the classical virtual machine for classical computing are located at the IaaS layer, making quantum virtual machines for quantum computing and
  • the communication between classical virtual machines for classical computing becomes intra-cluster communication, which reduces the delay of cross-cluster communication, reduces communication overhead and data delay, and exerts the advantages of quantum computing.
  • the SaaS (Software-as-a-service, software as a service) layer mainly provides application scenario solutions for users, and specifically, mainly provides user interfaces for user services, so that through The user interface acquires the quantum and classical hybrid programming language corresponding to the to-be-executed task.
  • the SaaS layer includes: a user programming module, configured to provide a user interface, so as to obtain the quantum and classical hybrid programming language corresponding to the task to be executed through the user interface.
  • the classical programming language can be python
  • the quantum programming can be the graphical quantum circuit programming
  • the quantum circuit can be embedded in python to form a quantum and classical hybrid programming language, which is very convenient for users to use.
  • the PaaS (Platform as a Service, platform as a service) layer is mainly an efficient task division and resource scheduling platform, and the PaaS layer mainly includes: a quantum and classical algorithm compilation module, used for the quantum and classical algorithm compilation module.
  • the hybrid programming language performs algorithm compilation and task separation to obtain quantum computing tasks and classical computing tasks corresponding to the tasks to be executed; a resource management and scheduling module is used to allocate resources for the quantum computing tasks and the classical computing tasks respectively.
  • the resource management and scheduling module is used for: calculating the first resource to be allocated corresponding to the quantum computing task, and according to the first resource to be allocated, from the IaaS (Infrastructure as a Service, infrastructure as a service, infrastructure as a service) ) layer of idle classical servers to determine the first target classical server, so as to deploy quantum virtual machine on the first target classical server; calculate the second resource to be allocated corresponding to the classical computing task, and according to the first target classical server Second, the resources to be allocated determine a second target classic server from idle classic servers in the IaaS layer, so as to use the second classic server to perform the classic computing task.
  • IaaS Infrastructure as a Service, infrastructure as a service, infrastructure as a service
  • the resource management and scheduling module will determine how many classical servers are needed for deploying quantum virtual machines according to the quantum computing task, and then can determine the corresponding number of the first target classical servers from the idle classical servers in the IaaS layer, Used to deploy quantum virtual machines.
  • the resource management and scheduling module also determines how many classic servers are needed for classic computing according to classic computing tasks, and then can determine a corresponding number of second target classic servers from idle classic servers in the IaaS layer for execution.
  • Classic computing tasks will determine how many classical servers are needed for deploying quantum virtual machines according to the quantum computing task, and then can determine the corresponding number of the first target classical servers from the idle classical servers in the IaaS layer, Used to deploy quantum virtual machines.
  • the resource management and scheduling module also determines how many classic servers are needed for classic computing according to classic computing tasks, and then can determine a corresponding number of second target classic servers from idle classic servers in the IaaS layer for execution.
  • Classic computing tasks will determine how many classical servers are needed for deploying quantum virtual machines according to the quantum computing task
  • the PaaS layer includes: a quantum virtual machine deployment module, used to obtain the information of the first target classical server determined by the resource management and scheduling module, and deploy the quantum virtual machine on the first target classical server machine. That is, the PaaS layer further includes a quantum virtual machine deployment module, and after the resource management and scheduling module allocates resources, installs the quantum virtual machine on the first target classical server that needs to install the quantum virtual machine.
  • the PaaS layer further includes a cloud platform operating system.
  • the IaaS layer mainly requires complete infrastructure construction.
  • the IaaS layer includes: a quantum virtual machine on the first target classical server for executing the quantum computing task; and a second target classical server for executing the classical computing task. Wherein, the first target classic server and the second target classic server need to be physically isolated.
  • the quantum virtual machine is deployed on some classical servers isolated according to the user's quantum computing task requirements.
  • the quantum virtual machine can provide quantum computing services.
  • For the user it is not aware of whether the task is running on a physical quantum computer or a quantum virtual machine. of.
  • Classical computing tasks are performed by other classical servers in the IaaS layer, so that the communication between quantum and classical is within a cluster, and the delay is greatly reduced.
  • an embodiment of the present application discloses a specific quantum and classical hybrid cloud platform, which includes:
  • the user programming module 111 in the SaaS layer 11 is used to provide a user interface, so as to obtain the quantum and classical hybrid programming language corresponding to the task to be performed through the user interface;
  • the solution provision module 112 in the SaaS layer 11 is used to provide machine vision solutions and reinforcement learning solutions;
  • the quantum and classical algorithm compilation module 121 in the PaaS layer 12 is used to perform algorithm compilation and task separation on the quantum and classical hybrid programming language, and obtain quantum computing tasks and classical computing tasks corresponding to the tasks to be executed;
  • the resource management and scheduling module 122 in the PaaS layer 12 is used to allocate resources for the quantum computing tasks and the classical computing tasks respectively;
  • the quantum virtual machine deployment module 123 in the PaaS layer 12 is used to obtain the information of the first target classical server determined by the resource management and scheduling module, and deploy the quantum virtual machine on the first target classical server;
  • the quantum virtual machine 131 on the first target classical server in the IaaS layer 13 is used to execute the quantum computing task;
  • the storage device 133 in the IaaS layer 13 is used for data storage
  • the network device 134 in the IaaS layer 13 is used for communication between different devices in the IaaS layer;
  • the infrastructure management module 135 in the IaaS layer 13 is used to manage, monitor and operate the infrastructure of the IaaS layer.
  • the SaaS layer in addition to the aforementioned user programming module 111, which is used to provide a user interface, so as to obtain the quantum and classical hybrid programming language corresponding to the task to be executed through the user interface, it also includes: Solution providing module 112 for providing machine vision solutions and reinforcement learning solutions.
  • the SaaS layer also provides solutions for some scenarios.
  • Machine vision is one of the core directions in the field of AI (Artificial Intelligence), and is widely used in object recognition, object detection, pixel-level semantic segmentation, etc.
  • AI Artificial Intelligence
  • the SaaS layer provides a set of quantum convolutional neural network solutions, using quantum revolving gates and quantum controlled NOT gates to build fully linear quantum (convolutional) neural networks, with Strong generalization performance.
  • the SaaS layer can provide several object recognition-oriented quantum convolutional neural network models based on the cloud platform.
  • Classical reinforcement learning has the disadvantage of poor learning effect in complex scenarios, while quantum reinforcement learning has a large available environment space and behavior space due to strong quantum parallelism, and the speed of obtaining the optimal solution is far faster than that of classical reinforcement learning.
  • the SaaS layer can provide several quantum reinforcement learning solutions for typical scenarios.
  • the IaaS layer includes, in addition to the aforementioned quantum virtual machine 131 on the first target classical server, for executing the quantum computing task, and the second target classical server 132, for executing one of the classical computing tasks
  • it also includes: a storage device 133 for data storage; a network device 134 for communication between different devices in the IaaS layer; and an infrastructure management module 135 for the infrastructure of the IaaS layer. Facility management, monitoring and maintenance.
  • the IaaS layer further includes a storage device 133 for data storage, a network device 134 for communication between different devices in the IaaS layer, and management of the basic settings in the IaaS layer , Infrastructure management module 135 for monitoring and operation and maintenance.
  • the infrastructure management module 135 monitors the occupancy and remaining situation of resources in real time, and feeds it back to the PaaS layer for task evaluation. It also needs to perform hardware fault detection and automatic repair. When automatic repair is not possible, an early warning is issued and the operation and maintenance personnel perform manual repair.
  • the occupied resources can be released, so that the released resources can be calculated into idle resources for subsequent tasks to call.
  • FIG. 3 is a schematic diagram of a quantum and classical hybrid cloud platform. It includes a SaaS layer, a PaaS layer and an IaaS layer, wherein the SaaS layer includes a user programming module for providing a user interface, so as to obtain the quantum and classical hybrid programming language corresponding to the task to be executed through the user interface.
  • the classical programming language supports python, and quantum programming can be used to program graphical quantum circuits. Quantum circuits can be embedded in python to form a quantum and classical hybrid programming language, which is very convenient for users to use.
  • the SaaS layer also includes solution providing modules for providing machine vision solutions and reinforcement learning solutions.
  • the PaaS layer includes: a quantum and classical algorithm compilation module, which is used to perform algorithm compilation and task separation on the quantum and classical hybrid programming language to obtain quantum computing tasks and classical computing tasks corresponding to the tasks to be executed; a resource management and scheduling module , which is used to allocate resources for the quantum computing tasks and the classical computing tasks respectively; the quantum virtual machine deployment module is used to obtain the information of the first target classical server determined by the resource management and scheduling module. Deploy a quantum virtual machine on the target classical server.
  • the PaaS layer also includes the cloud platform operating system.
  • the PaaS layer includes a classical server, and the classical server includes a first target classical server for deploying a quantum virtual machine, and the second target classical server for executing the classical computing task.
  • the PaaS layer further includes: a quantum virtual machine on the first target classical server for executing the quantum computing task.
  • the PaaS layer also includes: a storage device for data storage; a network device for communication between different devices in the IaaS layer; and an infrastructure management module (that is, the infrastructure management, Monitoring and operation and maintenance), used to manage, monitor and operate the infrastructure of the IaaS layer.
  • an infrastructure management module that is, the infrastructure management, Monitoring and operation and maintenance
  • an embodiment of the present application discloses a specific quantum and classical hybrid task execution method, which is applied to the aforementioned quantum and classical hybrid cloud platform.
  • the method includes:
  • Step S11 Obtain the quantum and classical hybrid programming language corresponding to the task to be executed through the user interface on the SaaS layer.
  • the quantum and classical hybrid programming language corresponding to the task to be executed through the user interface on the SaaS layer.
  • the classical programming language supported by the user interface can be python
  • the quantum programming can be graphical quantum circuit programming
  • the quantum circuit can be embedded in python to form a quantum and classical hybrid programming language, which is very convenient for users to use.
  • Step S12 Perform algorithm compilation and task separation on the quantum and classical hybrid programming language through the PaaS layer to obtain quantum computing tasks and classical computing tasks corresponding to the tasks to be executed, which are respectively the quantum computing tasks and the classical computing tasks resource allocation.
  • the task to be executed is divided into the quantum computing task and the classical computing task, it is necessary to determine how many classical servers are needed for deploying quantum virtual machines according to the quantum computing task, and then the idle time of the IaaS layer can be obtained.
  • a corresponding number of first target classical servers are determined in the classical servers for deploying quantum virtual machines.
  • Step S13 The quantum virtual machine is used to perform the quantum computing task and the classical server is used to perform the classical computing task through the IaaS layer according to the resource allocation situation of the PaaS layer.
  • the IaaS layer After the resource allocation is performed, the IaaS layer also needs to use the quantum virtual machine to perform the quantum computing task and use the classical server to perform the classical computing task according to the situation of resource allocation by the PaaS layer. That is, the quantum computing task is performed by the quantum virtual machine in the IaaS layer that has been deployed on the first target classical server, and the classical computing task is performed by the second target classical server in the IaaS layer.
  • quantum computing tasks and classical computing tasks can be processed synchronously, realizing the simultaneous and fast execution of dual computing modes, and maximizing the utilization of computing resources, which improves the task processing efficiency, and the communication between quantum computing and classical computing is an internal part of the IaaS layer. Communication within the cluster reduces communication overhead and data latency.
  • the occupied resources may be released, so that the released resources can be calculated into free resources for subsequent tasks to be called.
  • the SaaS layer also provides solutions for some scenarios.
  • Machine vision is one of the core directions in the field of AI (Artificial Intelligence), and is widely used in object recognition, object detection, pixel-level semantic segmentation, etc.
  • AI Artificial Intelligence
  • the SaaS layer provides a set of quantum convolutional neural network solutions, using quantum revolving gates and quantum controlled NOT gates to build fully linear quantum (convolutional) neural networks, with Strong generalization performance.
  • the SaaS layer may provide several object recognition-oriented quantum convolutional neural network models based on the cloud platform.
  • Classical reinforcement learning has the disadvantage of poor learning effect in complex scenarios, while quantum reinforcement learning has a large available environment space and behavior space due to strong quantum parallelism, and the speed of obtaining the optimal solution is far faster than that of classical reinforcement learning.
  • the SaaS layer can provide several quantum reinforcement learning solutions for typical scenarios.
  • users can train quantum convolutional neural networks through the machine vision solutions and reinforcement learning solutions provided by the SaaS layer, so as to use the trained quantum convolutional neural networks for object recognition, object detection, and pixel-level semantic segmentation. Wait.
  • a software module can be placed in random access memory (RAM), internal memory, read only memory (ROM), electrically programmable ROM, electrically erasable programmable ROM, registers, hard disk, removable disk, CD-ROM, or any other in the technical field. in any other known form of storage medium.
  • RAM random access memory
  • ROM read only memory
  • electrically programmable ROM electrically erasable programmable ROM
  • registers hard disk, removable disk, CD-ROM, or any other in the technical field. in any other known form of storage medium.

Landscapes

  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • Software Systems (AREA)
  • General Physics & Mathematics (AREA)
  • Physics & Mathematics (AREA)
  • General Engineering & Computer Science (AREA)
  • Condensed Matter Physics & Semiconductors (AREA)
  • Evolutionary Computation (AREA)
  • Mathematical Analysis (AREA)
  • Mathematical Optimization (AREA)
  • Pure & Applied Mathematics (AREA)
  • Computing Systems (AREA)
  • Data Mining & Analysis (AREA)
  • Mathematical Physics (AREA)
  • Computational Mathematics (AREA)
  • Artificial Intelligence (AREA)
  • Optical Modulation, Optical Deflection, Nonlinear Optics, Optical Demodulation, Optical Logic Elements (AREA)
  • Stored Programmes (AREA)

Abstract

一种量子与经典混合云平台以及任务执行方法，该云平台包括：SaaS层，用于提供用户接口，以便通过所述用户接口获取待执行任务对应的量子与经典混合编程语言；PaaS层，用于对所述量子与经典混合编程语言进行算法编译和任务分离，得到所述待执行任务对应的量子计算任务和经典计算任务，并分别为所述量子计算任务和经典计算任务分配资源；IaaS层，用于根据所述PaaS层分配资源的情况利用量子虚拟机执行所述量子计算任务和利用经典服务器执行所述经典计算任务。这样能够减小通信开销以及数据延迟，提高任务处理效率，发挥量子计算优势。

Description

一种量子与经典混合云平台以及任务执行方法
本申请要求在2020年11月19日提交中国专利局、申请号为202011301939.3、发明名称为“一种量子与经典混合云平台以及任务执行方法”的中国专利申请的优先权，其全部内容通过引用结合在本申请中。
技术领域
本申请涉及云计算技术领域，特别涉及一种量子与经典混合云平台以及任务执行方法。
背景技术
随着人工智能、大数据、物联网等新一代信息技术的发展，当今社会已经迈入万物互联时代，数据正在成为信息领域最大的资源。然而，数据量的爆炸式增长对传统计算系统的算力提出了巨大挑战，如何针对海量数据进行快速有效处理是近年来限制机器学习、大数据、量子化学、新药物研发等技术进一步实用化的主要障碍。其难题主要有二：(1)由于摩尔定律即将走向极限，电子芯片算力无法通过提升工艺来进行算力的提升；(2)目前内存墙的限制越来越严重，电子芯片受到极大的约束。
量子计算是最有希望解决以上问题的途径之一。而量子云平台将成为量子计算未来长期主要的存在形式。目前量子计算云平台主要是单纯提供在线的量子芯片或仿真的服务，然而在目前的量子计算云平台中需要在本地或其它设置中完成经典计算，然后再与量子计算云平台之间进行频繁的通信以完成整个计算，由于需要进行经典计算集群和量子计算集群之间的频繁通信，跨集群通信造成大量的通信开销，由此造成量子芯片和经典设备之间的数据延迟过大，甚至丧失原本的量子优势。
发明内容
有鉴于此，本申请的目的在于提供一种量子与经典混合云平台以及任务执行方法，能够减小通信开销以及数据延迟，提高任务处理效率，发挥量子计算优势。其具体方案如下：
第一方面，本申请公开了一种量子与经典混合云平台，包括：
SaaS层，用于提供用户接口，以便通过所述用户接口获取待执行任务对应的量子与经典混合编程语言；
PaaS层，用于对所述量子与经典混合编程语言进行算法编译和任务分离，得到所述待执行任务对应的量子计算任务和经典计算任务，并分别为所述量子计算任务和经典计算任务分配资源；
IaaS层，用于根据所述PaaS层分配资源的情况利用量子虚拟机执行所述量子计算任务和利用经典服务器执行所述经典计算任务。
可选地，所述SaaS层，包括：
用户编程模块，用于提供用户接口，以便通过所述用户接口获取待执行任务对应的量子与经典混合编程语言。
可选地，所述资源管理与调度模块，用于：
计算所述量子计算任务对应的第一待分配资源，并根据所述第一待分配资源从所述IaaS层中空闲的经典服务器中确定出第一目标经典服务器，以便在所述第一目标经典服务器上部署量子虚拟机；
计算所述经典计算任务对应的第二待分配资源，并根据所述第二待分配资源从所述IaaS层中空闲的经典服务器中确定出第二目标经典服务器，以便利用所述第二经典服务器执行所述经典计算任务。
可选地，所述PaaS层，包括：
量子虚拟机部署模块，用于获取所述资源管理与调度模块确定出的第一目标经典服务器的信息，并在所述第一目标经典服务器上部署量子虚拟机。
可选地，所述IaaS层，包括：
所述第一目标经典服务器上的量子虚拟机，用于执行所述量子计算任务；
第二目标经典服务器，用于执行所述经典计算任务。
可选地，所述IaaS层，包括：
存储设备，用于进行数据存储；
网络设备，用于进行所述IaaS层中不同设备之间的通信。
可选地，所述IaaS层，包括：
基础设施管理模块，用于对所述IaaS层的基础设施进行管理、监控与运维。
可选地，所述SaaS层，包括：
解决方案提供模块，用于提供机器视觉解决方案和强化学习解决方案。
第二方面，本申请公开了一种量子与经典混合任务执行方法，应用于前述的量子与经典混合云平台，包括：
通过SaaS层上的用户接口获取待执行任务对应的量子与经典混合编程语言；
通过PaaS层对所述量子与经典混合编程语言进行算法编译和任务分离，得到所述待执行任务对应的量子计算任务和经典计算任务，并分别为所述量子计算任务和经典计算任务分配资源；
通过IaaS层根据所述PaaS层分配资源的情况利用量子虚拟机执行所述量子计算任务和利用经典服务器执行所述经典计算任务。
可见，本申请公开了一种量子与经典混合云平台，该云平台包括SaaS层，用于提供用户接口，以便通过所述用户接口获取待执行任务对应的量子与经典混合编程语言；PaaS层，用于对所述量子与经典混合编程语言进行算法编译和任务分离，得到所述待执行任务对应的量子计算任务和经典计算任务，并分别为所述量子计算任务和经典计算任务分配资源；以及IaaS层，用于根据所述PaaS层分配资源的情况利用量子虚拟机执行所述量子计算任务和利用经典服务器执行所述经典计算任务。由此，本申请在SaaS层通过用户接口，以便用户可以通过所述用户接口输入量子与经典混合编程语言，解决了现有量子云平台中只支持量子编程的单一模式带来的用户不友好问题，且在所述PaaS层进行量子与经典混合编程语言的编译时，便将待执行任务分割成量子计算任务和经典计算任务，并配置对应的IaaS层资源执行对应的任务，实现了双计算模式同步快速执行，且最大化地利用了计算资源，提高了任务处理效率，此外，进行量子计算的量子虚拟机和进行经典计算的经典虚拟机都位于IaaS层，使得进行量子计算的量子虚拟机和进行经典计算的经典虚拟机之间的通信成为集群内通信，减少了跨集群通信的时延，减小了通信开销以及数据延迟，发挥了量子计算优势。
附图说明
为了更清楚地说明本申请实施例或现有技术中的技术方案，下面将对实 施例或现有技术描述中所需要使用的附图作简单地介绍，显而易见地，下面描述中的附图仅仅是本申请的实施例，对于本领域普通技术人员来讲，在不付出创造性劳动的前提下，还可以根据提供的附图获得其他的附图。
图1为本申请公开的一种量子与经典混合云平台结构示意图；
图2为本申请公开的一种具体的量子与经典混合云平台结构示意图；
图3为本申请公开的一种具体的量子与经典混合云平台结构示意图；
图4为本申请公开的一种具体的量子与经典混合任务执行方法流程图。
具体实施方式
下面将结合本申请实施例中的附图，对本申请实施例中的技术方案进行清楚、完整地描述，显然，所描述的实施例仅仅是本申请一部分实施例，而不是全部的实施例。基于本申请中的实施例，本领域普通技术人员在没有做出创造性劳动前提下所获得的所有其他实施例，都属于本申请保护的范围。
目前，量子计算云平台主要是单纯提供在线的量子芯片或仿真的服务，然而在目前的量子云架构中需要在本地或其它设置中完成经典计算，然后再与量子计算云平台之间进行频繁的通信完成整个计算，由于需要进行经典计算集群和量子计算集群之间的频繁通信，跨集群通信造成大量的通信开销，由此造成量子芯片和经典设备之间的数据延迟过大，甚至丧失原本的量子优势。有鉴于此，本申请提出了一种量子与经典混合云平台，能够减小通信开销以及数据延迟，提高任务处理效率，发挥量子计算优势。
参见图1所示，本申请实施例公开了一种量子与经典混合云平台，该平台包括：
SaaS层11，用于提供用户接口，以便通过所述用户接口获取待执行任务对应的量子与经典混合编程语言；
PaaS层12，用于对所述量子与经典混合编程语言进行算法编译和任务分离，得到所述待执行任务对应的量子计算任务和经典计算任务，并分别为所述量子计算任务和经典计算任务分配资源；
IaaS层13，用于根据所述PaaS层分配资源的情况利用量子虚拟机执行所述量子计算任务和利用经典服务器执行所述经典计算任务。
可见，本申请公开了一种量子与经典混合云平台，该云平台包括SaaS层， 用于提供用户接口，以便通过所述用户接口获取待执行任务对应的量子与经典混合编程语言；PaaS层，用于对所述量子与经典混合编程语言进行算法编译和任务分离，得到所述待执行任务对应的量子计算任务和经典计算任务，并分别为所述量子计算任务和经典计算任务分配资源；以及IaaS层，用于根据所述PaaS层分配资源的情况利用量子虚拟机执行所述量子计算任务和利用经典服务器执行所述经典计算任务。由此，本申请在SaaS层通过用户接口，以便用户可以通过所述用户接口输入量子与经典混合编程语言，解决了现有量子云平台中只支持量子编程的单一模式带来的用户不友好问题，且在所述PaaS层进行量子与经典混合编程语言的编译时，便将待执行任务分割成量子计算任务和经典计算任务，并配置对应的IaaS层资源执行对应的任务，实现了双计算模式同步快速执行，且最大化地利用了计算资源，提高了任务处理效率，此外，进行量子计算的量子虚拟机和进行经典计算的经典虚拟机都位于IaaS层，使得进行量子计算的量子虚拟机和进行经典计算的经典虚拟机之间的通信成为集群内通信，减少了跨集群通信的时延，减小了通信开销以及数据延迟，发挥了量子计算优势。
在具体的实施过程中，所述SaaS(Software-as-a-service，软件即服务)层，主要是面向用户提供应用场景解决方案，具体的，主要是提供面向用户服务的用户接口，以便通过所述用户接口获取所述待执行任务对应的量子与经典混合编程语言。也即，所述SaaS层，包括：用户编程模块，用于提供用户接口，以便通过所述用户接口获取待执行任务对应的量子与经典混合编程语言。其中，经典编程语言可以为python，量子编程可以为图像化的量子线路编程，量子线路可嵌入到python中形成量子与经典混合编程语言，极大的方便用户使用。
具体的，所述PaaS(Platform as a Service，平台即服务)层主要是高效的任务划分和资源调度平台，所述PaaS层主要包括：量子与经典算法编译模块，用于对所述量子与经典混合编程语言进行算法编译和任务分离，得到所述待执行任务对应的量子计算任务和经典计算任务；资源管理与调度模块，用于分别为所述量子计算任务和经典计算任务分配资源。
其中，所述资源管理与调度模块，用于：计算所述量子计算任务对应的第一待分配资源，并根据所述第一待分配资源从所述IaaS(Infrastructure as a  Service，基础设施即服务)层中空闲的经典服务器中确定出第一目标经典服务器，以便在所述第一目标经典服务器上部署量子虚拟机；计算所述经典计算任务对应的第二待分配资源，并根据所述第二待分配资源从所述IaaS层中空闲的经典服务器中确定出第二目标经典服务器，以便利用所述第二经典服务器执行所述经典计算任务。
也即，所述资源管理与调度模块会根据量子计算任务确定出需要多少经典服务器用于部署量子虚拟机，然后便可以从IaaS层的空闲经典服务器中确定出相应数量的第一目标经典服务器，用于部署量子虚拟机。所述资源管理与调度模块也会根据经典计算任务确定出需要多少经典服务器用于进行经典计算，然后便可以从IaaS层的空闲经典服务器中确定出相应数量的第二目标经典服务器，用于执行经典计算任务。
相应的，所述PaaS层，包括：量子虚拟机部署模块，用于获取所述资源管理与调度模块确定出的第一目标经典服务器的信息，并在所述第一目标经典服务器上部署量子虚拟机。也即，所述PaaS层还包括量子虚拟机部署模块，在所述资源管理与调度模块分配好资源之后，在需要安装量子虚拟机的第一目标经典服务器上安装量子虚拟机。
此外，所述PaaS层还包括云平台操作系统。
在实际的实施过程中，所述IaaS层主要需要进行完备的基础设施构建。所述IaaS层，包括：所述第一目标经典服务器上的量子虚拟机，用于执行所述量子计算任务；第二目标经典服务器，用于执行所述经典计算任务。其中，所述第一目标经典服务器和所述第二目标经典服务器需要进行物理隔离。
其中，量子虚拟机部署在根据用户的量子计算任务需求隔离出的部分经典服务器上，量子虚拟机可提供量子计算服务，对于用户而言，不感知任务是运行在物理量子计算机还是量子虚拟机上的。通过IaaS层中的其他部分经典服务器执行经典计算任务，这样量子经典之间的通信是在一个集群内，延迟大大降低。
参见图2所示，本申请实施例公开了一种具体的量子与经典混合云平台，该平台包括：
SaaS层11中的用户编程模块111，用于提供用户接口，以便通过所述用 户接口获取待执行任务对应的量子与经典混合编程语言；
SaaS层11中的解决方案提供模块112，用于提供机器视觉解决方案和强化学习解决方案；
PaaS层12中的量子与经典算法编译模块121，用于对所述量子与经典混合编程语言进行算法编译和任务分离，得到所述待执行任务对应的量子计算任务和经典计算任务；
PaaS层12中的资源管理与调度模块122，用于分别为所述量子计算任务和经典计算任务分配资源；
PaaS层12中的量子虚拟机部署模块123，用于获取所述资源管理与调度模块确定出的第一目标经典服务器的信息，并在所述第一目标经典服务器上部署量子虚拟机；
IaaS层13中的第一目标经典服务器上的量子虚拟机131，用于执行所述量子计算任务；
IaaS层13中的第二目标经典服务器132，用于执行所述经典计算任务；
IaaS层13中的存储设备133，用于进行数据存储；
IaaS层13中的网络设备134，用于进行所述IaaS层中不同设备之间的通信；
IaaS层13中的基础设施管理模块135，用于对所述IaaS层的基础设施进行管理、监控与运维。
在具体的实施过程中，所述SaaS层上，除了前述的用户编程模块111，用于提供用户接口，以便通过所述用户接口获取待执行任务对应的量子与经典混合编程语言之外，还包括解决方案提供模块112，用于提供机器视觉解决方案和强化学习解决方案。
也即，所述SaaS层也提供面向部分场景的解决方案。其一，提供强泛化能力的机器视觉解决方案。机器视觉是AI(Artificial Intelligence，人工智能)领域的核心方向之一，广泛应用于物体识别、物体探测、像素级语义分割等。但传统卷积神经网络的过拟合现象严重，所述SaaS层提供一套量子卷积神经网络解决方案，利用量子旋转门和量子受控非门构建全线性量子(卷积)神经网络，具有强大泛化性能。所述SaaS层可以基于云平台提供数个面向物体识别的量子卷积神经网络模型。其二，提供面向复杂场景的量子强化学习解 决方案。经典强化学习在复杂场景下有学习效果差的缺点，而量子强化学习由于强大的量子并行性对应的可用环境空间和行为空间均很大，获取最优解的速度也远超经典强化学习。所述SaaS层可以提供数个面向典型场景的量子强化学习解决方案。
在实际应用中，所述IaaS层除了包括前述的第一目标经典服务器上的量子虚拟机131，用于执行所述量子计算任务，第二目标经典服务器132，用于执行所述经典计算任务之外，还包括：存储设备133，用于进行数据存储；网络设备134，用于进行所述IaaS层中不同设备之间的通信；以及基础设施管理模块135，用于对所述IaaS层的基础设施进行管理、监控与运维。
也即，所述IaaS层还包括用于进行数据存储的存储设备133，用于进行IaaS层中不同设备之间的通信的网络设备134，以及用于对所述IaaS层中的基础设置进行管理、监控与运维的基础设施管理模块135。
所述基础设施管理模块135实时监控资源的占用和剩余情况，反馈给PaaS层做任务评估，也需要进行硬件的故障检测和自动修复，无法自动修复时发出预警由运维人员进行手动修复。
在所述量子计算任务和所述经典计算任务结束之后，可以将占用的资源释放，以便将释放出来的资源计算到空闲资源中，以便后续的任务进行调用。
参见图3所示，为量子与经典混合云平台示意图。包括SaaS层、PaaS层和IaaS层，其中，SaaS层包括用户编程模块，用于提供用户接口，以便通过所述用户接口获取待执行任务对应的量子与经典混合编程语言。经典编程语言支持python，量子编程可以为图像化的量子线路编程，量子线路可嵌入到python中形成量子与经典混合编程语言，极大的方便用户使用。SaaS层还包括解决方案提供模块，用于提供机器视觉解决方案和强化学习解决方案。PaaS层包括：量子与经典算法编译模块，用于对所述量子与经典混合编程语言进行算法编译和任务分离，得到所述待执行任务对应的量子计算任务和经典计算任务；资源管理与调度模块，用于分别为所述量子计算任务和经典计算任务分配资源；量子虚拟机部署模块，用于获取所述资源管理与调度模块确定出的第一目标经典服务器的信息，并在所述第一目标经典服务器上部署量子虚拟机。PaaS层还包括云平台操作系统。PaaS层包括经典服务器，经典服务器包括第一目标经典服务器，用于部署量子虚拟机，以及所述第二目标经典 服务器，用于执行所述经典计算任务。所述PaaS层还包括：第一目标经典服务器上的量子虚拟机，用于执行所述量子计算任务。所述PaaS层还包括：存储设备，用于进行数据存储；网络设备，用于进行所述IaaS层中不同设备之间的通信；以及基础设施管理模块(也即，图中的基础设置管理、监控与运维)，用于对所述IaaS层的基础设施进行管理、监控与运维。
参见图4所示，本申请实施例公开了一种具体的量子与经典混合任务执行方法，应用于前述的量子与经典混合云平台，该方法包括：
步骤S11：通过SaaS层上的用户接口获取待执行任务对应的量子与经典混合编程语言。
首先需要通过SaaS层上的用户接口获取待执行任务对应的量子与经典混合编程语言。其中，所述用户接口支持的经典编程语言可以为python，量子编程可以为图像化的量子线路编程，量子线路可嵌入到python中形成量子与经典混合编程语言，极大的方便用户使用。
步骤S12：通过PaaS层对所述量子与经典混合编程语言进行算法编译和任务分离，得到所述待执行任务对应的量子计算任务和经典计算任务，并分别为所述量子计算任务和经典计算任务分配资源。
在获取所述量子与经典混合编程语言之后，还需要通过PaaS层对所述量子与经典混合编程语言进行算法编译和任务分离，将待执行任务分割成量子计算任务和经典计算任务，并分别为量子计算任务和经典计算任务分配资源。
也即，将所述待执行任务分割成所述量子计算任务和所述经典计算任务之后，需要根据量子计算任务确定出需要多少经典服务器用于部署量子虚拟机，然后便可以从IaaS层的空闲经典服务器中确定出相应数量的第一目标经典服务器，用于部署量子虚拟机。以及根据经典计算任务确定出需要多少经典服务器用于进行经典计算，然后便可以从IaaS层的空闲经典服务器中确定出相应数量的第二目标经典服务器，用于执行经典计算。
步骤S13：通过IaaS层根据所述PaaS层分配资源的情况利用量子虚拟机执行所述量子计算任务和利用经典服务器执行所述经典计算任务。
进行资源分配之后，还需要由IaaS层根据所述PaaS层分配资源的情况利用量子虚拟机执行所述量子计算任务和利用经典服务器执行所述经典计算任 务。也即，通过所述IaaS层中的已部署在第一目标经典服务器上的量子虚拟机执行所述量子计算任务，以及通过所述IaaS层中的第二目标经典服务器执行所述经典计算任务。这样量子计算任务和经典计算任务可以同步进行处理，实现双计算模式同步快速执行，以及计算资源最大化利用，提高了任务处理效率，且量子计算和经典计算之间的通信为IaaS层内部的一个集群之内的通信，减少了通信开销以及数据延迟。
在所述待执行任务对应的量子计算任务和所述经典计算任务结束之后，可以将占用的资源释放，以便将释放出来的资源计算到空闲资源中，以便后续的任务进行调用。
在实际应用中，所述SaaS层也提供面向部分场景的解决方案。其一，提供强泛化能力的机器视觉解决方案。机器视觉是AI(Artificial Intelligence，人工智能)领域的核心方向之一，广泛应用于物体识别、物体探测、像素级语义分割等。但传统卷积神经网络的过拟合现象严重，所述SaaS层提供一套量子卷积神经网络解决方案，利用量子旋转门和量子受控非门构建全线性量子(卷积)神经网络，具有强大泛化性能。所述SaaS层可以基于云平台将提供数个面向物体识别的量子卷积神经网络模型。其二，提供面向复杂场景的量子强化学习解决方案。经典强化学习在复杂场景下有学习效果差的缺点，而量子强化学习由于强大的量子并行性对应的可用环境空间和行为空间均很大，获取最优解的速度也远超经典强化学习。所述SaaS层可以提供数个面向典型场景的量子强化学习解决方案。
所以用户可以通过所述SaaS层提供的机器视觉解决方案和强化学习解决方案进行量子卷积神经网络的训练等，以便利用训练后的量子卷积神经网络进行物体识别、物体探测、像素级语义分割等。
此外，所述SaaS层、所述PaaS层和所述IaaS层中还可以进行其他的处理操作，具体可以参考前述实施例中公开的内容，在此不再进行赘述。
本说明书中各个实施例采用递进的方式描述，每个实施例重点说明的都是与其它实施例的不同之处，各个实施例之间相同或相似部分互相参见即可。对于实施例公开的装置而言，由于其与实施例公开的方法相对应，所以描述的比较简单，相关之处参见方法部分说明即可。
结合本文中所公开的实施例描述的方法或算法的步骤可以直接用硬件、处理器执行的软件模块，或者二者的结合来实施。软件模块可以置于随机存储器(RAM)、内存、只读存储器(ROM)、电可编程ROM、电可擦除可编程ROM、寄存器、硬盘、可移动磁盘、CD-ROM、或技术领域内所公知的任意其它形式的存储介质中。
最后，还需要说明的是，在本文中，诸如第一和第二之类的关系术语仅仅用来将一个实体或者操作与另一个实体或者操作区分开来，而不一定要求或者暗示这些实体或操作之间存在任何这种实际的关系或者顺序。而且，术语“包括”、“包含”或者其任何其他变体意在涵盖非排他性的包含，从而使得一系列包含其他要素的过程、方法、物品或者设备不仅包括那些要素，而且还包括没有明确列出的其他要素，或者是还包括为这种过程、方法、物品或者设备所固有的要素。在没有更多限制的情况下，由语句“包括一个……”限定的要素，并不排除在包括所述要素的过程、方法、物品或者设备中还存在另外的相同要素。
以上对本申请所提供的一种量子与经典混合云平台以及量子与经典混合任务执行方法进行了详细介绍，本文中应用了具体个例对本申请的原理及实施方式进行了阐述，以上实施例的说明只是用于帮助理解本申请的方法及其核心思想；同时，对于本领域的一般技术人员，依据本申请的思想，在具体实施方式及应用范围上均会有改变之处，综上所述，本说明书内容不应理解为对本申请的限制。

Claims (10)

  1. 一种量子与经典混合云平台，其特征在于，包括：
    SaaS层，用于提供用户接口，以便通过所述用户接口获取待执行任务对应的量子与经典混合编程语言；
    PaaS层，用于对所述量子与经典混合编程语言进行算法编译和任务分离，得到所述待执行任务对应的量子计算任务和经典计算任务，并分别为所述量子计算任务和经典计算任务分配资源；
    IaaS层，用于根据所述PaaS层分配资源的情况利用量子虚拟机执行所述量子计算任务和利用经典服务器执行所述经典计算任务。
  2. 根据权利要求1所述的量子与经典混合云平台，其特征在于，所述SaaS层，包括：
    用户编程模块，用于提供用户接口，以便通过所述用户接口获取待执行任务对应的量子与经典混合编程语言。
  3. 根据权利要求1所述的量子与经典混合云平台，其特征在于，所述PaaS层，包括：
    量子与经典算法编译模块，用于对所述量子与经典混合编程语言进行算法编译和任务分离，得到所述待执行任务对应的量子计算任务和经典计算任务；
    资源管理与调度模块，用于分别为所述量子计算任务和经典计算任务分配资源。
  4. 根据权利要求3所述的量子与经典混合云平台，其特征在于，所述资源管理与调度模块，用于：
    计算所述量子计算任务对应的第一待分配资源，并根据所述第一待分配资源从所述IaaS层中空闲的经典服务器中确定出第一目标经典服务器，以便在所述第一目标经典服务器上部署量子虚拟机；
    计算所述经典计算任务对应的第二待分配资源，并根据所述第二待分配资源从所述IaaS层中空闲的经典服务器中确定出第二目标经典服务器，以便利用所述第二经典服务器执行所述经典计算任务。
  5. 根据权利要求4所述的量子与经典混合云平台，其特征在于，所述PaaS层，包括：
    量子虚拟机部署模块，用于获取所述资源管理与调度模块确定出的第一 目标经典服务器的信息，并在所述第一目标经典服务器上部署量子虚拟机。
  6. 根据权利要求4所述的量子与经典混合云平台，其特征在于，所述IaaS层，包括：
    所述第一目标经典服务器上的量子虚拟机，用于执行所述量子计算任务；
    所述第二目标经典服务器，用于执行所述经典计算任务。
  7. 根据权利要求1所述的量子与经典混合云平台，其特征在于，所述IaaS层，包括：
    存储设备，用于进行数据存储；
    网络设备，用于进行所述IaaS层中不同设备之间的通信。
  8. 根据权利要求1所述的量子与经典混合云平台，其特征在于，所述IaaS层，包括：
    基础设施管理模块，用于对所述IaaS层的基础设施进行管理、监控与运维。
  9. 根据权利要求1至8任一项所述的量子与经典混合云平台，其特征在于，所述SaaS层，包括：
    解决方案提供模块，用于提供机器视觉解决方案和强化学习解决方案。
  10. 一种量子与经典混合任务执行方法，其特征在于，应用于权利要求1至9任一项所述的量子与经典混合云平台，包括：
    通过SaaS层上的用户接口获取待执行任务对应的量子与经典混合编程语言；
    通过PaaS层对所述量子与经典混合编程语言进行算法编译和任务分离，得到所述待执行任务对应的量子计算任务和经典计算任务，并分别为所述量子计算任务和经典计算任务分配资源；
    通过IaaS层根据所述PaaS层分配资源的情况利用量子虚拟机执行所述量子计算任务和利用经典服务器执行所述经典计算任务。
PCT/CN2021/121221 2020-11-19 2021-09-28 一种量子与经典混合云平台以及任务执行方法 WO2022105440A1 (zh)

Priority Applications (1)

Application Number Priority Date Filing Date Title
US18/020,616 US20230297401A1 (en) 2020-11-19 2021-09-28 Hybrid quantum-classical cloud platform and task execution method

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN202011301939.3A CN112465146B (zh) 2020-11-19 2020-11-19 一种量子与经典混合云平台以及任务执行方法
CN202011301939.3 2020-11-19

Publications (1)

Publication Number Publication Date
WO2022105440A1 true WO2022105440A1 (zh) 2022-05-27

Family

ID=74836749

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2021/121221 WO2022105440A1 (zh) 2020-11-19 2021-09-28 一种量子与经典混合云平台以及任务执行方法

Country Status (3)

Country Link
US (1) US20230297401A1 (zh)
CN (1) CN112465146B (zh)
WO (1) WO2022105440A1 (zh)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN116566844A (zh) * 2023-07-06 2023-08-08 湖南马栏山视频先进技术研究院有限公司 一种基于多云融合的数据管控方法与多云融合管理平台

Families Citing this family (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN112465146B (zh) * 2020-11-19 2022-05-24 苏州浪潮智能科技有限公司 一种量子与经典混合云平台以及任务执行方法
CN112862104B (zh) * 2021-04-01 2024-02-27 中国科学技术大学 混合型量子计算机架构及其执行计算任务的方法
EP4332841A1 (en) * 2021-06-23 2024-03-06 Origin Quantum Computing Technology (Hefei) Co., Ltd Method, system and apparatus for processing quantum computing task, and operating system
CN113902120A (zh) * 2021-09-18 2022-01-07 中国人民解放军战略支援部队信息工程大学 异构云解算平台混合计算任务动态自适应划分调度方法及系统
CN114048857B (zh) * 2021-10-22 2024-04-09 天工量信(苏州)科技发展有限公司 算力分配方法、装置及算力服务器

Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN109213603A (zh) * 2018-05-31 2019-01-15 合肥本源量子计算科技有限责任公司 一种用于对接量子计算机与用户的云平台及平台运行方法
CN110083454A (zh) * 2019-05-05 2019-08-02 济南浪潮高新科技投资发展有限公司 一种带有量子计算机的混合云服务编排方法
CN112465146A (zh) * 2020-11-19 2021-03-09 苏州浪潮智能科技有限公司 一种量子与经典混合云平台以及任务执行方法

Family Cites Families (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US9537953B1 (en) * 2016-06-13 2017-01-03 1Qb Information Technologies Inc. Methods and systems for quantum ready computations on the cloud
CN108874538B (zh) * 2018-05-31 2020-10-13 合肥本源量子计算科技有限责任公司 用于调度量子计算机的调度服务器、调度方法及应用方法
US10705883B2 (en) * 2018-06-19 2020-07-07 Microsoft Technology Licensing, Llc Dynamic hybrid computing environment

Patent Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN109213603A (zh) * 2018-05-31 2019-01-15 合肥本源量子计算科技有限责任公司 一种用于对接量子计算机与用户的云平台及平台运行方法
CN110083454A (zh) * 2019-05-05 2019-08-02 济南浪潮高新科技投资发展有限公司 一种带有量子计算机的混合云服务编排方法
CN112465146A (zh) * 2020-11-19 2021-03-09 苏州浪潮智能科技有限公司 一种量子与经典混合云平台以及任务执行方法

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
PETER J. KARALEKAS; NIKOLAS A. TEZAK; ERIC C. PETERSON; COLM A. RYAN; MARCUS P. DA SILVA; ROBERT S. SMITH: "A quantum-classical cloud platform optimized for variational hybrid algorithms", ARXIV.ORG, 31 May 2020 (2020-05-31), 201 Olin Library Cornell University Ithaca, NY 14853 , XP081660137, DOI: 10.1088/2058-9565/ab7559 *

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN116566844A (zh) * 2023-07-06 2023-08-08 湖南马栏山视频先进技术研究院有限公司 一种基于多云融合的数据管控方法与多云融合管理平台
CN116566844B (zh) * 2023-07-06 2023-09-05 湖南马栏山视频先进技术研究院有限公司 一种基于多云融合的数据管控方法与多云融合管理平台

Also Published As

Publication number Publication date
US20230297401A1 (en) 2023-09-21
CN112465146B (zh) 2022-05-24
CN112465146A (zh) 2021-03-09

Similar Documents

Publication Publication Date Title
WO2022105440A1 (zh) 一种量子与经典混合云平台以及任务执行方法
US10896064B2 (en) Coordinated, topology-aware CPU-GPU-memory scheduling for containerized workloads
WO2022262167A1 (zh) 集群资源调度方法及装置、电子设备和存储介质
US11809913B2 (en) Load balancing of cloned virtual machines
US9405582B2 (en) Dynamic parallel distributed job configuration in a shared-resource environment
US20180239642A1 (en) Resource Configuration System, Resource Configuration Method and Resource Configuration Program
EP3067797A1 (en) Creating new cloud resource instruction set architecture
US20200193964A1 (en) Method and device for training an acoustic model
US20210191759A1 (en) Elastic Execution of Machine Learning Workloads Using Application Based Profiling
US11232009B2 (en) Model-based key performance indicator service for data analytics processing platforms
US11184263B1 (en) Intelligent serverless function scaling
CN110580195B (zh) 一种基于内存热插拔的内存分配方法和装置
CN107391256A (zh) 一种基于云计算技术的继电保护定值整定计算架构及方法
US11831706B1 (en) System and method for distributed management of storage systems based on intent
Huang et al. Design and implementation of an edge computing platform architecture using Docker and Kubernetes for machine learning
US10579419B2 (en) Data analysis in storage system
CN112463296A (zh) 一种创建虚拟机的方法及装置
KR102320324B1 (ko) 쿠버네티스 환경에서의 이종 하드웨어 가속기 활용 방법 및 이를 이용한 장치
CN116389542A (zh) 具有可配置的池化资源的平台
US20240143377A1 (en) Overlay container storage driver for microservice workloads
CN112527450B (zh) 基于不同资源的超融合自适应方法、终端及系统
CN112433812A (zh) 一种虚拟机跨集群迁移方法、系统、设备及计算机介质
CN107402807A (zh) 在计算机系统中有效提升多任务执行效率的方法、系统和处理器
US20220326986A1 (en) Scheduling workloads on partitioned resources of a host system in a container-orchestration system
CN109558214B (zh) 异构环境下宿主机资源管理方法、装置和存储介质

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 21893591

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 21893591

Country of ref document: EP

Kind code of ref document: A1