Quantum state data storage method and apparatus, and device, medium and quantum simulation system

By constructing a high-performance and low-cost heterogeneous hierarchical storage architecture, the problem of high storage costs in quantum simulation systems is solved, achieving both cost reduction and performance assurance.

WO2026238057A1PCT designated stage Publication Date: 2026-11-19BEIJING SURSEN NETWORK TECH
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
PCT/CN2026/096987
Authority / Receiving Office
WO · WO
Patent Type
Applications
Current Assignee / Owner
Priority Date
2025-05-15
Filing Date
2026-05-14
Publication Date
2026-11-19

AI Technical Summary

Technical Problem

In existing quantum simulation systems, the storage and transmission of quantum state data increases exponentially with scale, leading to excessively high storage costs.

Method used

A heterogeneous hierarchical storage architecture is constructed, consisting of a high-performance first storage area and a low-cost, high-capacity second storage area. The high-performance storage area temporarily stores key quantum state data, while the low-cost storage area persistently stores all quantum state data, thus optimizing the data transmission path.

Benefits of technology

Significantly reduce storage costs while ensuring system performance and real-time performance, control network latency and storage costs within a reasonable range, and achieve a balance between system performance and storage costs.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN2026096987_19112026_PF_FP_ABST
    Figure CN2026096987_19112026_PF_FP_ABST
Patent Text Reader

Abstract

Disclosed in the present application are a quantum state data storage method and apparatus, and a device, a medium and a quantum simulation system. The method comprises: transmitting part of quantum state data stored in a second storage area to a first storage area (S210); at least one computing unit executing a simulation quantum computing task on the basis of the quantum state data in the first storage area (S220); and storing in the second storage area the quantum state data in the first storage area after the simulation quantum computing task is executed (S230). A data storage path is optimized by constructing a heterogeneous hierarchical storage architecture comprising a high-performance first storage area and a low-cost and high-capacity second storage area, and quantum state data related to a current simulation quantum computing task is temporarily stored by means of the high-performance first storage area, so that high-speed access to key quantum state data is realized, and all quantum state data is stored by means of the second storage area, thereby significantly reducing storage costs.
Need to check novelty before this filing date? Find Prior Art

Description

Quantum state data storage methods, devices, equipment, media, and quantum simulation systems

[0001] This application claims priority to Chinese Patent Application No. 202510629779.1, filed on May 15, 2025, entitled "Quantum State Data Storage Method, Apparatus, Device, Medium and Quantum Simulation System", the entire contents of which are incorporated herein by reference. Technical Field

[0002] This application relates to the field of quantum simulation, specifically to a quantum state data storage method, device, equipment, medium, and quantum simulation system. Background Technology

[0003] Quantum simulation can use quantum technology to simulate or study complex quantum systems. It is a means of simulating and studying quantum computing. Due to the characteristics of quantum mechanics, in the process of simulating quantum computing, it is necessary to traverse the entire state vector, that is, to store and manipulate the state vector of all quantum states.

[0004] In traditional quantum simulation systems, all quantum state data is typically stored in high-performance storage media such as DRAM or even HBM to meet the requirements for quantum state data storage and transmission, ensuring system performance. However, as the scale of quantum simulation systems increases, the amount of quantum state data that needs to be stored and transmitted increases exponentially, leading to exponentially higher storage costs. Therefore, reducing the storage cost of quantum simulation systems has become an urgent problem to be solved. Summary of the Invention

[0005] In view of this, embodiments of this application provide a quantum state data storage method, apparatus, device, medium, and quantum simulation system. By constructing a storage system including a high-performance first storage area and a high-capacity second storage area, data storage costs are optimized, and storage costs are reduced while ensuring the performance of the quantum simulation system.

[0006] In a first aspect, embodiments of this application provide a quantum state data storage method applied to a quantum simulation system. The quantum simulation system includes at least one computing unit, a first storage area, and a second storage area. The performance of the first storage area is higher than that of the second storage area, and the storage space of the second storage area is larger than that of the first storage area. The second storage area is used to store quantum state data. The method includes: transferring a portion of the quantum state data stored in the second storage area to the first storage area; at least one computing unit performing a simulated quantum computing task based on the quantum state data in the first storage area; and storing the quantum state data generated after performing the simulated quantum computing task in the first storage area to the second storage area.

[0007] In one possible implementation, transferring a portion of the quantum state data stored in the second storage area to the first storage area includes: determining first quantum state data for each computing unit based on the simulated quantum computing task currently to be performed by each computing unit in at least one computing unit, wherein the first quantum state data is quantum state data related to the currently performed simulated quantum computing task; and transferring the first quantum state data stored in the second storage area to the first storage area.

[0008] In one possible implementation, the first storage area is divided into at least two sub-storage areas, the at least two sub-storage areas including a first sub-storage area and a second sub-storage area; transferring the first quantum state data stored in the second storage area to the first storage area includes: transferring the first quantum state data to the first sub-storage area, causing at least one computing unit to execute a current simulated quantum computing task based on the first quantum state data in the first sub-storage area; while at least one computing unit is executing the current simulated quantum computing task based on the first quantum state data in the first sub-storage area, transferring the second quantum state data stored in the second storage area to the second sub-storage area, wherein the second quantum state data is quantum state data related to the next simulated computing task to be executed after at least one computing unit has completed the current simulated quantum computing task.

[0009] In one possible implementation, the method further includes: dividing the quantum state data stored in the second storage area into multiple data blocks; wherein transferring a portion of the quantum state data stored in the second storage area to the first storage area includes: transferring a portion of the quantum state data stored in the second storage area to the first storage area in units of data blocks; wherein storing the quantum state data after performing the simulated quantum computing task in the first storage area to the second storage area includes: storing the quantum state data after performing the simulated quantum computing task in the first storage area to the second storage area in units of data blocks.

[0010] In one possible implementation, the method further includes: adding an index to each of the multiple data blocks, the index being used to record the storage location of each data block in the second storage area; determining the storage location of the data block containing the first quantum state data in the second storage area based on the index, wherein transferring a portion of the quantum state data stored in the second storage area to the first storage area on a data block basis includes: transferring the data block stored at the storage location to the first storage area.

[0011] In one possible implementation, transferring a portion of the quantum state data stored in the second storage area to the first storage area includes: transferring the portion of the quantum state data stored in the second storage area to the first storage area via DMA.

[0012] In one possible implementation, the first storage area is a multi-port storage area, so that the data transmission process does not need to go through the CPU.

[0013] In one possible implementation, the first storage area is any one or a combination of high-bandwidth memory (HBM), dynamic random access memory (DRAM), static random access memory (SRAM), magnetoresistive random access memory (MRAM), ferroelectric random access memory (FeRAM), phase-change memory (PCM), 3D XPoint, and the internal storage of the computing unit.

[0014] In one possible implementation, the second storage area is any one or a combination of solid-state drives (SSDs), hard disk drives (HDDs), resistive random access memory (RRAM), magnetoresistive random access memory (MRAM), ferroelectric random access memory (FeRAM), phase-change memory (PCM), 3DXPoint, NAND Flash, and NOR Flash.

[0015] In one possible implementation, at least one computing unit includes any one or any combination of a CPU, a CPU core, a GPU, a GPU core, a DPU, a DPU core, a TPU, a TPU core, a virtual machine, a container, a virtual core, an FPGA, and an FPGA core.

[0016] In one possible implementation, at least one computing unit includes at least two sets of computing units, wherein each set of computing units includes at least one computing unit, and the first storage area includes at least two physically separate sub-storage areas, with each of the at least two sets of computing units corresponding one-to-one with the at least two sub-storage areas included in the first storage area.

[0017] In one possible implementation, the second storage area is mounted as at least one block device to the server where at least one computing unit resides.

[0018] In one possible implementation, the block device is a shared block device.

[0019] In one possible implementation, simulating a quantum computing task involves performing specified computational operations on a specified range of quantum state data.

[0020] In one possible implementation, a quantum circuit is divided into multiple groups of quantum gates, and a set of simulated quantum computing tasks is determined for each computing unit in at least one computing unit according to the type and parameters of each group of quantum gates, wherein a set of simulated quantum computing tasks consists of zero or at least one simulated quantum computing task.

[0021] In one possible implementation, the first storage area and the second storage area are connected via a storage network.

[0022] In one possible implementation, the storage network is any one or a combination of CXL, PCIe, SAS, IB, RoCE, and Ethernet storage networks.

[0023] Secondly, embodiments of this application provide a quantum simulation system, which includes: at least one computing unit, a first storage area and a second storage area, wherein the performance of the first storage area is higher than that of the second storage area, the storage space of the second storage area is larger than that of the first storage area, the second storage area is used to store quantum state data, and at least one computing unit is used to execute the quantum state data storage method of the first aspect.

[0024] Thirdly, embodiments of this application provide a quantum state data storage device applied to a quantum simulation system. The quantum simulation system includes at least one computing unit, a first storage area, and a second storage area. The performance of the first storage area is higher than that of the second storage area, and the storage space of the second storage area is larger than that of the first storage area. The second storage area is used to store quantum state data. The device includes: a data transmission module configured to transmit a portion of the quantum state data stored in the second storage area to the first storage area; a task execution module configured to have at least one computing unit perform a simulated quantum computing task based on the quantum state data in the first storage area; and a data storage module configured to store the quantum state data after the simulated quantum computing task is performed in the first storage area to the second storage area.

[0025] Fourthly, embodiments of this application provide an electronic device including at least one processor and a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, the instructions being executed by the at least one processor to enable the at least one processor to perform the quantum state data storage method of the first aspect.

[0026] Fifthly, embodiments of this application provide a computer-readable storage medium storing instructions that, when executed by a processor of an electronic device, enable the electronic device to perform the quantum state data storage method of the first aspect.

[0027] This application provides a quantum state data storage method, apparatus, device, medium, and quantum simulation system. The method optimizes the data transmission path by constructing a heterogeneous hierarchical storage architecture that includes a high-performance first storage area and a low-cost, high-capacity second storage area. Key quantum state data (relevant quantum state data required in the current quantum computing task) in the simulated quantum computing task are temporarily stored in the high-performance first storage area, enabling high-speed access to key quantum state data, i.e., efficient reading, and ensuring the real-time performance of the simulated quantum computing task. All quantum state data are stored in the low-cost, high-capacity second storage area, effectively reducing the overall storage cost and keeping network latency and storage costs within a reasonable range, thus achieving a balance between system performance and storage cost. Attached Figure Description

[0028] The accompanying drawings are provided to further illustrate the present application and form part of the specification. They are used together with the embodiments of the present application to explain the application and do not constitute a limitation thereof. The above and other features and advantages will become more apparent to those skilled in the art from the detailed example embodiments described with reference to the accompanying drawings, in which:

[0029] Figure 1 is a schematic diagram of the structure of a quantum simulation system provided in some embodiments of this application.

[0030] Figure 2 is an exemplary flowchart of a quantum state data storage method provided in some embodiments of this application.

[0031] Figure 3 is an exemplary flowchart of transferring data from a second storage area to a first storage area according to some embodiments of this application.

[0032] Figure 4 is a flowchart illustrating a method for storing quantum state data in a first storage area according to some embodiments of this application.

[0033] Figure 5 is a schematic diagram of the structure of the first storage area provided in some embodiments of this application.

[0034] Figure 6 is a system block diagram of a quantum state data storage device provided in some embodiments of this application.

[0035] Figure 7 is a schematic diagram of the structure of an electronic device provided in some embodiments of this application. Detailed Implementation

[0036] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0037] Application Overview

[0038] In existing quantum simulation systems, all quantum state data is typically stored using high-performance storage media to achieve high-speed access to the quantum state data and ensure system performance.

[0039] However, in the process of simulating quantum computing, quantum simulation systems need to traverse the entire state vector, that is, they need to store and manipulate the state vectors of all quantum states. As the scale of the quantum simulation system increases, the amount of quantum state data that needs to be stored and transmitted increases exponentially, requiring huge storage space to support quantum simulation operations.

[0040] Taking a quantum simulation system with n=40 qubits as an example, corresponding to 2 40 There are 2 complex numbers, each occupying 16 bytes, for a total data size of 2. 40 ×16 bytes (approximately 17.6TB), if all storage is done using DRAM (DRAM costs $50 / GB), the storage cost would be approximately $887,000. Assuming the configuration of the computing unit (GPU) and other costs (including network configuration) are approximately $20,000, the total cost would be $1,087,000. Therefore, the storage cost accounts for approximately 97% of the total cost of the quantum simulation system.

[0041] In other words, the storage architecture described above is used to store large-scale quantum state data, but the cost of the storage medium accounts for a high proportion of the total cost of the quantum simulation system, resulting in high storage costs.

[0042] To address the aforementioned issues, this application creatively proposes a heterogeneous hierarchical storage architecture that optimizes data transmission paths and reduces storage costs by constructing a high-performance first storage area and a low-cost, high-capacity second storage area. Specifically, the high-performance first storage area enables efficient storage and retrieval of key quantum state data (quantum state data related to the currently executed simulated quantum computing task) during quantum simulation computation, achieving efficient access to key quantum state data. Simultaneously, the low-cost, high-capacity second storage area enables persistent storage of all quantum state data in the quantum simulation system, significantly reducing the overall system storage cost and keeping network latency and storage costs within a reasonable range, thus ensuring system performance while significantly reducing storage costs.

[0043] Various non-limiting embodiments of this application will now be described in detail with reference to the accompanying drawings.

[0044] Exemplary quantum simulation system

[0045] This application provides a quantum simulation system 100. Figure 1 is a schematic diagram of the structure of a quantum simulation system provided in some embodiments of this application. As shown in Figure 1, the quantum simulation system 100 may include at least one computing unit 110, a first storage area 120, a storage network 130, and a second storage area 140. The performance of the first storage area 120 is higher than that of the second storage area 140, and the storage space of the second storage area 140 is larger than that of the first storage area 120. The first storage area 120 is used to temporarily store quantum state data related to the quantum computing task being performed or about to be performed by the at least one computing unit 110. The second storage area 140 is used to store (e.g., persistent storage) all the quantum state data in the quantum simulation system 100.

[0046] In some embodiments, the quantum simulation system 100 may include a computing unit 110, a first storage area 120, and a second storage area 140. In this case, the first storage area 120 may be a local disk, for example, the internal storage of the server where the computing unit 110 resides, such as a laptop hard drive, and the computing unit can directly access the first storage area 120 (i.e., the local disk). The first storage area 120 and the second storage area 140 may also be storage media in independent storage devices. In this case, the computing unit 110, the first storage area 120, and the second storage area 140 can be connected through a storage network to realize data transmission. For details, please refer to the following detailed description of Figure 1.

[0047] In some embodiments, the quantum simulation system 100 can be a distributed quantum simulation system, at least one computing unit 110 includes at least one set of computing units (two sets are used as an example in the figure), wherein each set of computing units includes at least one computing unit, the first storage area 120 includes at least one sub-storage area, and the first storage area 120 and the second storage area 140 are connected through a storage network 130.

[0048] Specifically, the distributed quantum simulation system can consist of multiple network zones. For example, a multi-network-zone quantum simulation system may include at least two network zones and a storage network 130. The storage network 130 may include a storage network access network 131 and a storage network aggregation network 132. Each network zone includes at least one sub-storage area of ​​a first storage area 120, a second storage area 140, a storage network access network 131, and optional computing units 110. For example, the first storage area 120 may be the memory of a server containing at least one computing unit 110, such as DRAM, and its sub-storage areas are the memory of each server. Each sub-storage area of ​​the first storage area 120 includes at least one storage medium, and the second storage area includes at least one storage medium. Sub-storage areas of the first storage area within the same network zone are connected to the second storage area 140 via the storage network access network 131, and the storage network access networks 131 of different network zones are interconnected via the storage network aggregation network 132. The first storage area 120 includes two sub-storage areas that are physically separated from each other, and the two computing units 110 correspond one-to-one with the two sub-storage areas.

[0049] It should be noted that the two computing units 110 shown in Figure 1, the number of the two sub-storage areas included in the first storage area 120, and the type of storage network 130 are merely examples and are not specifically limited here.

[0050] The computing unit is the core unit for performing simulated quantum computing tasks. For example, the computing unit can be a computing node in a server. Specifically, the computing node can be a central processing unit (CPU), a CPU core, a graphics processing unit (GPU), a GPU core, a data processing unit (DPU), a DPU core, a TPU core, an FPGA core, or other hardware with computing capabilities. It can also be a virtual machine, container, process, or thread.

[0051] In some embodiments, the quantum simulation system 100 further includes at least one storage unit. Exemplarily, the storage unit may be a storage node in a server that includes a computing unit; specifically, the storage node may be a software module for implementing storage functionality.

[0052] In some embodiments, storage network 130 may be a high-performance switch storage network. For example, the switch storage network may be any one or a combination of CXL (Compute Express Link), PCIe (PCI Express), SAS (Serial Attached SCSI), IB (InfiniBand), RoCE (RDMA over Converged Ethernet), and Ethernet storage networks.

[0053] In some embodiments, the number of at least two sub-storage areas included in the first storage area can be flexibly set according to the amount of quantum state data that needs to be temporarily stored in the simulated quantum computing task and the processing speed of the computing unit simulating the quantum computing task. No specific limitation is made here.

[0054] The first storage area may consist of at least one high-performance storage medium. For example, the first storage area may be any one or a combination of high-bandwidth memory (HBM), dynamic random access memory (DRAM), static random access memory (SRAM), magnetoresistive random access memory (MRAM), ferroelectric random access memory (FeRAM), phase-change memory (PCM), 3D XPoint, and the internal storage of at least one computing unit.

[0055] The second storage area may consist of at least one storage medium with performance lower than that of the first storage area to reduce storage costs. For example, the second storage area may be any one or a combination of solid-state drives (SSDs), hard disk drives (HDDs), resistive random access memory (RRAM), magnetoresistive random access memory (MRAM), ferroelectric random access memory (FeRAM), phase-change memory (PCM), 3D XPoint, NAND Flash, and NOR Flash.

[0056] It should be noted that in practical applications, one or more suitable types of storage media can be selected as the first storage area and the second storage area according to the actual needs of the quantum simulation system. As long as the performance of the storage media in the first storage area is higher than that in the second storage area, the cost of the storage media in the second storage area is lower than that in the first storage area, and the total storage space of the second storage area is greater than that of the first storage area, no specific restrictions are imposed here.

[0057] Compared to the storage architecture in traditional quantum simulation systems, by constructing a heterogeneous hierarchical storage architecture that includes a high-performance first storage area and a low-cost, high-capacity second storage area, the proportion of high-performance storage areas in quantum simulation systems can be significantly reduced. Consequently, storage costs are also significantly reduced. At the same time, the high-performance first storage area can enable high-speed access to quantum state data related to the simulated quantum computing task, ensuring system performance.

[0058] Taking a quantum simulation system with 40 qubits (n=40) as an example, this corresponds to 240 complex numbers, each occupying 16 bytes, for a total data volume of 240 × 16 bytes (approximately 17.6 TB). In traditional storage architectures, if all storage is done using DRAM (with a cost of $50 / GB), the storage cost would be approximately $887,000. However, the heterogeneous tiered storage architecture proposed in this application can significantly reduce the number of high-performance storage media. Assuming the number of DRAMs is 1% of that in a traditional architecture, the storage cost would be $8,870. The remaining low-performance storage media would be SSDs, with a storage cost of $3,200. Therefore, the total storage cost in this application's solution is $12,070. Compared to traditional storage architectures, the cost of storage media is only 13% of that in traditional quantum simulation systems. The larger the scale of the quantum simulation system, the more significant the reduction in storage media cost will be due to this application's storage architecture.

[0059] In some embodiments, the first storage area and the second storage area are directly connected without needing to be connected through a storage network.

[0060] In some embodiments, the first storage area and the second storage area may be located on the same server. For example, the first storage area is the server's memory, and the second storage area is the server's hard disk.

[0061] In some embodiments, both the first storage area and the second storage area include multiple sub-storage areas, which correspond one-to-one with each other.

[0062] In some embodiments, the sub-storage areas of the first storage area and the sub-storage areas of the second storage area are physically separated.

[0063] In some embodiments, a sub-storage area of ​​the first storage area and its corresponding sub-storage area of ​​the second storage area are located on the same server.

[0064] In some embodiments, the computing units can be divided into multiple groups, with each group of computing units corresponding to a sub-storage area of ​​a first storage area and a sub-storage area of ​​a second storage area.

[0065] In some embodiments, a set of computing units and their corresponding sub-storage areas of the first storage area and the second storage area are all located on the same server.

[0066] In some embodiments, servers are connected via a storage network.

[0067] In some embodiments, the storage network connecting the server is any combination of one or more of Ethernet, IB, and RoCE.

[0068] In some embodiments, when a server (and the computing unit within that server) accesses quantum state data in a sub-storage area of ​​a second storage area located on another server, it needs to go through that other server.

[0069] In some embodiments, the storage media for the first storage area and / or the second storage area may be located in a storage device, such as a JBOD / JBOF / e-JBOF. The specific arrangement of the storage media can be determined based on its type and function in the system, and is not specifically limited here.

[0070] In some embodiments, the first storage area may be a dual-port or multi-port storage area so that the data transfer process does not need to go through the CPU.

[0071] In some embodiments, the second storage area is mounted to the server where at least one computing unit resides as a block device. For example, the shared storage pool can be mounted to the server as one or more block devices using protocols such as SCSI (Small Computer System Interface), iSCSI (Internet Small Computer System Interface), rbd (Rados Block Device), FC (Fiber Channel), NVME, NVMEoF (NVME over Fabrics), ATA, ATAPI, and FireWire. Each server can mount one or more block devices. The quantum simulation software on each server can directly read quantum state data as a block device, or it can create a file system on the block device to read quantum state data as a file.

[0072] In some embodiments, the block device can be a shared block device, that is, multiple servers share the same block device.

[0073] In practical applications, external factors or internal system failures may cause the physical link between the computing unit and the first storage area, or between the first storage area and the second storage area, to be broken. This would affect the computing unit's access to the data stored in the first storage area or the first storage area's access to the data stored in the second storage area, thus impacting the overall system performance. To avoid this, in some embodiments, as shown in Figure 1, at least two storage networks are connected between each first storage area and each second storage area. That is, at least two data transmission channels exist between each first storage area and each second storage area, ensuring that there are two independent physical links between any first storage area and any second storage area. If any link is broken, the first storage area can access the second storage area through the other link, giving the quantum simulation system good fault tolerance and ensuring high system availability.

[0074] In some embodiments, the quantum simulation system may further include a management module (not shown) for managing devices or resources or centralization within the system. Specifically, it can manage and allocate a first storage area, group quantum gates according to quantum circuits, divide a quantum circuit into multiple groups of quantum gates, and formulate a quantum simulation computation plan for each computing unit in at least one computing unit based on each group of quantum gates, including the quantum simulation computation tasks required by the computing unit when simulating the group of quantum gates.

[0075] This application further provides a quantum state data storage method. By constructing a heterogeneous hierarchical storage architecture comprising a high-performance first storage area and a low-cost, high-capacity second storage area, the data storage method is optimized, keeping network latency and storage costs within a reasonable range. Specifically, the high-performance first storage area temporarily stores quantum state data related to the current simulated quantum computing task, enabling high-speed access to quantum state data during simulated quantum computing and ensuring the real-time performance of the simulation task. Simultaneously, the low-cost, high-capacity second storage area persistently stores all quantum state data, significantly reducing overall storage costs and keeping network latency and storage costs within a reasonable range, thus achieving a balance between system performance and storage costs.

[0076] In some embodiments, the aforementioned quantum state data storage method can be integrated into at least one computing unit 110, a first storage area 120, and a second storage area 140 of the quantum simulation system 100 in the form of instructions, so that the computing unit can be used to execute the quantum state data storage method provided in this application.

[0077] Exemplary quantum state data storage method

[0078] To further illustrate the process of calling and transmitting quantum state data by the computing unit in the quantum simulation system shown in Figure 1 during a simulated quantum computing task, this application also provides an exemplary flowchart (Figure 2) of a quantum state data storage method. This quantum state data storage method can be applied, for example, to the quantum simulation system 100 shown in Figure 1, or to other systems with corresponding processing capabilities.

[0079] In some embodiments, the process shown in FIG2 can be executed by a computing unit in a quantum simulation system 100.

[0080] As shown in Figure 2, the method may include the following steps:

[0081] S210. Transfer some of the quantum state data stored in the second storage area to the first storage area.

[0082] The first storage region can be a storage area consisting of at least one storage region composed of high-performance storage media, used for temporary storage of quantum state data related to the current quantum computing task. For details regarding the first storage region, please refer to the relevant description in Figure 1.

[0083] Correspondingly, the second storage area can be a storage region consisting of at least one storage area composed of lower-performance (low-cost) storage media, used to store all quantum state data.

[0084] In some embodiments, the second storage area does not store all quantum state data, but rather stores a portion (typically most) of the quantum state data, with the remaining quantum state data stored in the first storage area or other storage areas (if any). For details regarding the contents of the first and second storage areas, please refer to the relevant description in Figure 1. The storage space of the first storage area is smaller than that of the second storage area.

[0085] Considering the potentially large scale of quantum state data, the storage space of the first storage area may be insufficient to store all of it. Therefore, a portion of the quantum state data (the quantum state data related to the simulated quantum computing task currently being executed by the computing unit) can be transferred to the first storage area, allowing the computing unit to execute the current simulated quantum computing task based on this portion of quantum state data. By using a first storage area and a second storage area with different performance and costs in combination to hierarchically store quantum state data in simulated quantum computing tasks, the need for high-performance storage media can be effectively reduced, storage media costs can be significantly lowered, and system performance can still be guaranteed.

[0086] Specifically, in some embodiments, as shown in FIG3, step S210 may include the following sub-steps:

[0087] S211. Determine the first quantum state data of each computing unit based on the simulated quantum computing task that each computing unit currently needs to perform in at least one computing unit.

[0088] S212. Transfer the first quantum state data stored in the second storage area to the first storage area.

[0089] The first quantum state data is the quantum state data related to the current simulated quantum computing task.

[0090] In some embodiments, the first storage area is further divided into at least two sub-storage areas, for example, it can be divided into a first sub-storage area and a second sub-storage area.

[0091] In some embodiments, the first quantum state data can be stored in the first sub-memory area. This allows the computing unit to execute the current simulated quantum computing task based on the first quantum state data in the first sub-memory area, while simultaneously storing the second quantum state data related to the next simulated quantum computing task to be executed in the second sub-memory area. Based on this, the computing unit can directly perform calculations on the second quantum state data after completing the calculation on the first quantum state data, reducing the idle waiting time of the computing unit and improving the performance of the simulation calculation. For details, please refer to the relevant descriptions in Figures 4 and 5.

[0092] S220, at least one computing unit performs a simulated quantum computing task based on quantum state data in the first storage area.

[0093] Simulated quantum computing tasks can be used to simulate the corresponding quantum computing process for computing units based on initial quantum state data.

[0094] In some embodiments, simulating a quantum computing task can be simulating the operation of at least one quantum gate on quantum state data. Specifically, it can simulate the state of a qubit and its quantum gate operations, or it can simulate complex quantum algorithms and quantum circuits, without being specifically limited to this.

[0095] S230. Store the quantum state data generated after performing the simulated quantum computing task in the first storage area to the second storage area.

[0096] Repeat the data transmission process of step S210, the execution process of step S220, and the data storage process of step S230 until all the quantum computing tasks to be performed by the computing unit have been completed and all quantum state data in the simulated quantum computing tasks have been stored in the second storage area.

[0097] In some embodiments, at least one computing unit includes at least two groups of computing units, and each group of computing units includes at least one computing unit, meaning that the quantum simulation task of the entire quantum simulation system is completed through the cooperation of multiple computing units. In this case, in step S210, it is necessary to first determine the current simulated quantum computing task of each computing unit in the at least two computing units, and then determine the first quantum state data of each computing unit based on the currently executed simulated quantum computing task, before executing steps S220 and S230.

[0098] In some embodiments, a simulation task plan can be formulated for each of at least two computing units based on the type and parameters of at least one quantum gate, wherein the simulation task plan can be understood as a set of quantum simulation computing tasks that the computing unit needs to perform.

[0099] In some embodiments, each computing unit is responsible for computing a certain range of quantum state data. Based on the size of the sub-storage area of ​​the first storage area corresponding to a computing unit, each computing task of the computing unit computes a portion of the quantum state data, and a group of computing tasks completes the computation of the quantum state data within the range that the computing unit is responsible for.

[0100] In some embodiments, a storage network can be used to connect the first storage area and the second storage area to establish a data transmission channel between any sub-storage area in the first storage area and the second storage area, so that the first storage area can read quantum state data (first quantum state data) related to the current simulated quantum computing task from the second storage area through the storage network.

[0101] In some embodiments, the storage network can be a high-performance switch storage network. For details about the storage network, please refer to the relevant description in Figure 1, which will not be repeated here.

[0102] In some embodiments, a quantum circuit can be divided into multiple groups of quantum gates. Based on the type and parameters of each group of quantum gates, a simulation task plan is formulated for each computing unit. The simulation task plan can be understood as a set of quantum simulation computing tasks that the computing unit needs to perform. For details, please refer to the relevant description in Figure 1.

[0103] Considering that some computing units may perform simulated quantum computing tasks while others are idle, in some embodiments, the simulation task plan may include zero or at least one simulation operation. The at least one simulation operation includes the quantum state data range corresponding to the simulation operation and the quantum computing operation to be simulated (i.e., the simulated quantum computing task). Based on the simulation task plan, each computing unit completes the simulation calculation of the specified quantum state data.

[0104] In some embodiments, to improve the overall efficiency of the data transmission process, before executing step S210, an index can be added to each of the multiple data blocks in the second storage area to record the storage location of each data block in the second storage area. For example, the quantum state data can be divided into multiple data blocks based on the continuity of the state vectors in the quantum state data stored in the second storage area. Specifically, it can be divided into multiple data blocks that contain the same number of quantum state data (i.e., the same size). Further, an index is added to each data block based on a lookup table structure, establishing a mapping relationship between each data table and the index. This allows the location of the quantum state data to be read directly through an index alignment strategy when reading quantum state data from the second storage area, reducing cross-node communication and improving the locality and accessibility of data access.

[0105] Considering that the scale of quantum state data to be transmitted increases accordingly in large-scale quantum simulation systems, frequent data read and write operations are required between the computing unit and the storage device. To further reduce unnecessary data read and write operations, in some embodiments, before executing step S210, the quantum state data stored in the second storage area can be pre-divided into multiple data blocks. Specifically, the quantum state data stored in the second storage area can be divided according to the continuity of the state vector in the quantum state data and the storage space size of each sub-storage area in the first storage area. For example, assume the size of the data block is S. block Let represent the number of bytes in each storage block. Then, the quantum state data can be divided according to the following formula:

[0106] Where, N block s represents the total number of data blocks into which the state vector of the entire quantum state data is divided. total S represents the storage size of the state vectors representing the total quantum state data stored in the second storage area. state The storage size occupied by the amplitude is n, where n represents the number of qubits in the simulated quantum system.

[0107] In some embodiments, the data transfer process of transferring the first quantum state data stored in the second storage area to the first storage area in step S210, the process of executing the simulated quantum computing task in step S220, and the data storage process of storing the quantum state data after executing the simulated quantum computing task in the first storage area to the second storage area in step S230 are all performed in units of data blocks. That is, step S210 may include transferring a portion of the quantum state data stored in the second storage area to the first storage area in units of data blocks, and step S230 may include storing the quantum state data after executing the simulated quantum computing task in the first storage area to the second storage area in units of data blocks, thereby realizing "block-level cache management", significantly reducing the reading overhead of irrelevant data and improving the reading efficiency of quantum state data.

[0108] In some embodiments, an index can be added to each of the multiple data blocks that have been divided. The index is used to record the storage location of each data block in the second storage area. For example, transferring a portion of the quantum state data stored in the second storage area to the first storage area in units of data blocks may include transferring the data block stored at the storage location to the first storage area.

[0109] In some embodiments, the storage location of the data block containing the first quantum state data in the second storage area can be determined based on the index of the quantum state data. Specifically, data blocks can be read from or written to the second storage area based on the index of the quantum state data.

[0110] In some embodiments, the data throughput of each data transmission link can be dynamically adjusted based on the dual-buffering mechanism and the real-time bandwidth of the storage network. For example, the data throughput of the transmission process of quantum state data from the second storage area to the first storage area and the transmission process of quantum state data from the first storage area to the computing unit can be adjusted to keep the data throughput of each data transmission link consistent.

[0111] In some embodiments, a portion of the quantum state data (first quantum state data) stored in the second storage area is directly transferred to the first storage area via DMA.

[0112] In some embodiments, the first storage area is a multi-port storage area (including dual-port and multi-port storage areas), which is equipped with independent read / write control lines and data transmission channels, so that the data transmission process does not need to go through the CPU.

[0113] Based on the aforementioned heterogeneous hierarchical storage architecture, which includes a high-performance first storage area and a low-cost, high-capacity second storage area, data storage paths are optimized to keep storage network latency and storage costs within a reasonable range. Specifically, the high-performance first storage area temporarily stores quantum state data that requires high-frequency access and is related to the current quantum computing simulation task, ensuring efficient storage and retrieval of quantum state data, i.e., the real-time performance of the simulation task. Simultaneously, the introduction of the low-cost, high-capacity second storage area stores all quantum state data in the system, effectively reducing overall storage costs and keeping storage network latency and storage costs within a reasonable range, thus achieving a balance between system performance and storage costs. Furthermore, while ensuring a balance between system performance and economy, the overall system's computational efficiency and resource utilization are significantly improved, providing strong support for high-load quantum simulation tasks.

[0114] Exemplary method for storing quantum state data in a first storage area

[0115] To further illustrate the process of storing quantum state data in the first storage area and the storage structure of the first storage area, embodiments of this application provide a flowchart (Figure 4) of a method for storing quantum state data in the first storage area and a structural diagram (Figure 5) of the first storage area.

[0116] In some embodiments, the process shown in FIG4 can be executed by a computing unit in a quantum simulation system 100.

[0117] As shown in Figure 4, the method may include the following steps:

[0118] S310. Transmit the first quantum state data to the first sub-storage area, so that at least one computing unit performs the current simulated quantum computing task based on the first quantum state data in the first sub-storage area.

[0119] S320. While at least one computing unit is performing the current simulated quantum computing task based on the first quantum state data in the first sub-storage area, the second quantum state data stored in the second storage area is transferred to the second sub-storage area.

[0120] The first quantum state data is quantum state data related to the current simulated quantum computing task, and the second quantum state data is quantum state data related to the next simulated quantum computing task to be executed after the computing unit finishes the current simulated quantum computing task.

[0121] In some embodiments, to further improve the accessibility of the storage area, as shown in FIG5, each sub-storage area 121 in the first storage area 120 is further divided into a first sub-storage area and a second sub-storage area. Based on the correlation between the quantum state data and the currently executed sub-simulated quantum computing task in the simulated quantum computing task, they are stored respectively in different sub-storage areas of the first storage area's sub-storage area. For example, each computing unit executes a sub-simulated quantum computing task (also called a sub-simulated computing task) within the simulated quantum computing task, and each computing unit corresponds to a sub-storage area in the first storage area. In this case, the quantum state data related to a sub-simulated quantum computing task corresponding to a computing unit can be stored in the sub-storage area corresponding to that computing unit.

[0122] Specifically, the quantum state data (first quantum state data) related to the currently executed simulated quantum computing task is stored in the first sub-storage area of ​​the first storage area, so that the computing unit can directly load the data block related to the current sub-simulation computing task from the first sub-storage area and execute the current sub-simulation quantum computing process based on the data block, thereby reducing unnecessary disk accesses and improving data access efficiency.

[0123] For example, a first sub-storage area of ​​capacity M can be pre-allocated within the first storage area to load only quantum state data relevant to the current simulated quantum computing task, thereby reducing unnecessary disk accesses. Assume the number of state vectors that can be cached in a single operation is:

[0124] Here, it indicates that the B state vector can be loaded into memory at most simultaneously during the calculation process.

[0125] In some embodiments, steps S310 and S320 can be executed simultaneously, that is, storing the portion of the quantum state data related to the currently executed sub-simulation computation process in the simulated quantum computing task in the first quantum state data into the first sub-storage area, and storing the quantum state data related to the sub-simulation computation process to be executed in the simulated quantum computing task in the first quantum state data into the second sub-storage area.

[0126] In some embodiments, while the computing unit is executing the current simulated quantum computing task based on the first quantum state data, the quantum state data related to the next simulated quantum computing task to be executed after the current simulated quantum computing task is stored in the second sub-storage area to ensure the continuity of the execution of the simulated quantum task.

[0127] In some embodiments, after the current sub-simulated quantum computing task is completed, the first quantum state data in the first sub-storage area is written to the second storage area, and the second quantum state data in the second sub-storage area is loaded into the first sub-storage area. The above process is repeated until the entire simulated quantum computing task corresponding to the computing unit is completed and all quantum states related to the entire simulated quantum computing task have been written to the second storage area.

[0128] Exemplary device

[0129] The methods and system embodiments of this application have been described in detail above with reference to Figures 1-5. The apparatus embodiments of this application are described in detail below. It should be understood that the descriptions of the method embodiments correspond to the descriptions of the apparatus embodiments; therefore, any parts not described in detail can be referred to the preceding method embodiments. Figure 6 is a schematic diagram of the system modules of a quantum state data storage device shown in some embodiments of this application.

[0130] As shown in Figure 6, the quantum state data storage device 600 may include a data transmission module 610, a task execution module 620, and a data storage module 630.

[0131] The data transmission module 610 can be configured to transmit a portion of the quantum state data stored in the second storage area to the first storage area.

[0132] The task execution module 620 can be configured to have at least one computing unit perform a simulated quantum computing task based on quantum state data in the first storage area.

[0133] The data storage module 630 can be configured to store the quantum state data after performing a simulated quantum computing task in the first storage area to the second storage area.

[0134] It should be understood that specific limitations regarding quantum state data storage devices can be found in the limitations regarding quantum state data storage methods described above, and will not be repeated here. Each module in the aforementioned quantum state data storage device can be implemented entirely or partially through software, hardware, or a combination thereof. These modules can be embedded in or independent of the processor in a computer device in hardware form, or stored in the memory of a computer device in software form, so that the processor can call and execute the operations corresponding to each module.

[0135] Exemplary electronic devices and storage media

[0136] This application also provides an electronic device, as shown in FIG7. The electronic device 700 provided by this application includes a memory 710, a processor 720, and an input / output interface 730. The memory 710, processor 720, and input / output interface 730 are connected via internal connection paths. The memory 710 is used to store instructions, and the processor 720 is used to execute the instructions stored in the memory 710 to control the input / output interface 730 to receive input data and information, and output operation results and other data.

[0137] It should be understood that in the embodiments of this application, the processor 720 may be a general-purpose central processing unit (CPU), GPU, FPGA, microprocessor, application-specific integrated circuit (ASIC), or one or more integrated circuits to execute related programs in order to implement the technical solutions provided in the embodiments of this application.

[0138] The memory 710 may include read-only memory and random access memory, and provides instructions and data to the processor 720. A portion of the processor 720 may also include non-volatile random access memory. For example, the processor 720 may also store device type information.

[0139] In implementation, each step of the above method can be completed by the integrated logic circuits in the hardware of the processor 720 or by instructions in software form. The quantum state data storage method disclosed in the embodiments of this application can be directly implemented by the hardware processor, or by a combination of hardware and software modules in the processor. The software modules can be located in random access memory, flash memory, read-only memory, programmable read-only memory, electrically erasable programmable memory, registers, or other mature storage media in the art. This storage medium is located in memory 710, and the processor 720 reads the information in memory 710 and completes the steps of the above method in combination with its hardware. To avoid repetition, it will not be described in detail here. This application also provides a computer program product, including a computer program / instructions. When the computer program / instruction processor in the computer program product provided in this application is executed, the quantum state data storage method provided in this application can be implemented.

[0140] All of the above-mentioned optional technical solutions can be combined in any way to form the optional embodiments of this application, and will not be described in detail here.

[0141] Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0142] It should be noted that in the apparatus, devices, and methods of this application, each module or step can be disassembled and / or recombined. These disassemblies and / or recombinations should be considered equivalent solutions of this application. The above description of the disclosed aspects is provided to enable any person skilled in the art to make or use this application. Various modifications to these aspects will be readily apparent to those skilled in the art, and the general principles defined herein can be applied to other aspects without departing from the scope of this application. Therefore, this application is not intended to be limited to the foregoing aspects, but rather to be accorded the widest scope consistent with the principles and novel features disclosed herein.

[0143] The above description is for illustrative purposes only, illustrating and describing the technical solutions of this application. Furthermore, this description is not intended to limit the embodiments of this application to the scope of the forms disclosed above. Although several exemplary aspects and embodiments have been discussed above, those skilled in the art can readily derive other variations, modifications, alterations, additions, and sub-combinations based on the above content.

[0144] It should be noted that in the description of this application, the terms "first," "second," "third," etc., are used for descriptive purposes only and should not be construed as indicating or implying relative importance. Furthermore, in the description of this application, unless otherwise stated, "a plurality of" means two or more.

[0145] The above are merely preferred embodiments of this application and are not intended to limit this application. Any modifications or equivalent substitutions made within the spirit and principles of this application should be included within the protection scope of this application.

Claims

1. A method of quantum state data storage, characterized by, The method is applied to a quantum simulation system, which includes at least one computing unit, a first storage area, and a second storage area. The performance of the first storage area is higher than that of the second storage area, and the storage space of the second storage area is larger than that of the first storage area. The second storage area is used to store quantum state data. The method includes: Transfer a portion of the quantum state data stored in the second storage area to the first storage area; The at least one computing unit performs a simulated quantum computing task based on the quantum state data in the first storage area; The quantum state data generated after performing the simulated quantum computing task in the first storage area is stored in the second storage area.

2. The quantum state data storage method of claim 1, wherein, The step of transferring a portion of the quantum state data stored in the second storage area to the first storage area includes: Based on the simulated quantum computing task that each computing unit in the at least one computing unit currently needs to execute, a first quantum state data of each computing unit is determined, wherein the first quantum state data is quantum state data related to the currently executed simulated quantum computing task; The first quantum state data stored in the second storage area is transferred to the first storage area.

3. The quantum state data storage method of claim 2, wherein, The first storage area is divided into at least two sub-storage areas, the at least two sub-storage areas including a first sub-storage area and a second sub-storage area; The step of transferring the first quantum state data stored in the second storage area to the first storage area includes: The first quantum state data is transmitted to the first sub-memory area, so that the at least one computing unit performs the current simulated quantum computing task based on the first quantum state data in the first sub-memory area; While the at least one computing unit is executing the current simulated quantum computing task based on the first quantum state data in the first sub-storage area, the second quantum state data stored in the second storage area is transferred to the second sub-storage area, wherein the second quantum state data is quantum state data related to the next simulated computing task to be executed after the at least one computing unit has completed the current simulated quantum computing task.

4. The quantum state data storage method of claim 2, wherein, The method further includes: The quantum state data stored in the second storage area is divided into multiple data blocks; The step of transferring a portion of the quantum state data stored in the second storage area to the first storage area includes: Transfer a portion of the quantum state data stored in the second storage area to the first storage area in units of data blocks; Wherein, storing the quantum state data generated after performing the simulated quantum computing task in the first storage area to the second storage area includes: The quantum state data generated after performing the simulated quantum computing task in the first storage area is stored in the second storage area in units of data blocks.

5. The quantum state data storage method of claim 4, wherein, The method further includes: An index is added to each of the plurality of data blocks, and the index is used to record the storage location of each data block in the second storage area; Based on the index, the storage location of the data block containing the first quantum state data in the second storage area is determined. The step of transferring a portion of the quantum state data stored in the second storage area to the first storage area in units of data blocks includes: The data block stored at the storage location is transferred to the first storage area.

6. The quantum state data storage method of any one of claims 1 to 5, wherein, The step of transferring a portion of the quantum state data stored in the second storage area to the first storage area includes: A portion of the quantum state data stored in the second storage area is transferred to the first storage area using DMA.

7. The quantum state data storage method of any one of claims 1 to 6, wherein, The first storage area is a multi-port storage area, so that the data transmission process does not need to go through the CPU.

8. The quantum state data storage method of any one of claims 1 to 7, wherein, The first storage area is any one or a combination of high-bandwidth memory (HBM), dynamic random access memory (DRAM), static random access memory (SRAM), magnetoresistive random access memory (MRAM), ferroelectric random access memory (FeRAM), phase change memory (PCM), 3D XPoint, and the internal storage of the computing unit.

9. The quantum state data storage method of any of claims 1 to 8, wherein, The second storage area is any one or a combination of solid-state drives (SSDs), hard disk drives (HDDs), resistive random access memory (RRAM), magnetoresistive random access memory (MRAM), ferroelectric random access memory (FeRAM), phase change memory (PCM), 3D XPoint, NAND Flash, and NOR Flash.

10. The quantum state data storage method of any one of claims 1 to 9, wherein, The at least one computing unit includes any one or any combination of a CPU, a CPU core, a GPU, a GPU core, a DPU, a DPU core, a TPU, a TPU core, a virtual machine, a container, a virtual core, an FPGA, and an FPGA core.

11. The quantum state data storage method of any one of claims 1 to 10, wherein, The at least one computing unit includes at least two sets of computing units, wherein each set of computing units includes at least one computing unit, and the first storage area includes at least two physically separate sub-storage areas, and the at least two sets of computing units correspond one-to-one with the at least two sub-storage areas included in the first storage area.

12. The quantum state data storage method of any one of claims 1 to 11, wherein, The second storage area is mounted to the server where the at least one computing unit is located in the manner of at least one block device.

13. The quantum state data storage method of claim 12, wherein, The block device is a shared block device.

14. The quantum state data storage method of any one of claims 1 to 13, wherein, The simulated quantum computing task includes performing specified computational operations on quantum state data within a specified range.

15. The quantum state data storage method of claim 14, wherein, The method further includes: A quantum circuit is divided into multiple groups of quantum gates. Based on the type and parameters of each group of quantum gates, a set of simulated quantum computing tasks is determined for each computing unit in the at least one computing unit. The set of simulated quantum computing tasks consists of zero or at least one simulated quantum computing task.

16. The quantum state data storage method of claim 14, wherein: The first storage area and the second storage area are connected via a storage network.

17. The quantum state data storage method of claim 16, wherein, The storage network is any one or a combination of CXL, PCIe, SAS, IB, RoCE and Ethernet storage networks.

18. A quantum simulation system, comprising: The system includes: at least one computing unit, a first storage area and a second storage area, wherein the performance of the first storage area is higher than that of the second storage area, the storage space of the second storage area is larger than that of the first storage area, the second storage area is used to store quantum state data, and the at least one computing unit is used to execute the quantum state data storage method according to any one of claims 1-17.

19. A quantum state data storage device, comprising: The device is applied to a quantum simulation system, the quantum simulation system including at least one computing unit, a first storage area and a second storage area, wherein the performance of the first storage area is higher than that of the second storage area, the storage space of the second storage area is larger than that of the first storage area, and the second storage area is used to store quantum state data. The device includes: The data transmission module is configured to transmit a portion of the quantum states stored in the second storage area to the first storage area; The task execution module is configured such that the at least one computing unit executes the simulated quantum computing task based on the quantum state data in the first storage area; The data storage module is configured to store the quantum state data after performing the simulated quantum computing task in the first storage area to the second storage area.

20. An electronic device, comprising: The electronic device includes: At least one processor; and A memory communicatively connected to the at least one processor; wherein, The memory stores instructions that can be executed by the at least one processor to enable the at least one processor to perform the quantum state data storage method according to any one of claims 1-17.

21. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores instructions that, when executed by the processor of the electronic device, enable the electronic device to perform the quantum state data storage method as described in any one of claims 1-17.