WO2021169386A1 - 一种图数据处理方法、装置、设备、介质 - Google Patents

一种图数据处理方法、装置、设备、介质 Download PDF

Info

Publication number
WO2021169386A1
WO2021169386A1 PCT/CN2020/126349 CN2020126349W WO2021169386A1 WO 2021169386 A1 WO2021169386 A1 WO 2021169386A1 CN 2020126349 W CN2020126349 W CN 2020126349W WO 2021169386 A1 WO2021169386 A1 WO 2021169386A1
Authority
WO
WIPO (PCT)
Prior art keywords
data
graph data
target
graph
flag value
Prior art date
Application number
PCT/CN2020/126349
Other languages
English (en)
French (fr)
Inventor
王媛丽
梅国强
王江为
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 US17/797,876 priority Critical patent/US20230334094A1/en
Publication of WO2021169386A1 publication Critical patent/WO2021169386A1/zh

Links

Images

Classifications

    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06FELECTRIC DIGITAL DATA PROCESSING
    • G06F16/00Information retrieval; Database structures therefor; File system structures therefor
    • G06F16/90Details of database functions independent of the retrieved data types
    • G06F16/901Indexing; Data structures therefor; Storage structures
    • G06F16/9024Graphs; Linked lists
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06FELECTRIC DIGITAL DATA PROCESSING
    • G06F16/00Information retrieval; Database structures therefor; File system structures therefor
    • G06F16/50Information retrieval; Database structures therefor; File system structures therefor of still image data
    • G06F16/51Indexing; Data structures therefor; Storage structures
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06FELECTRIC DIGITAL DATA PROCESSING
    • G06F15/00Digital computers in general; Data processing equipment in general
    • G06F15/76Architectures of general purpose stored program computers
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06TIMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T7/00Image analysis
    • G06T7/10Segmentation; Edge detection
    • G06T7/13Edge detection
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06NCOMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N5/00Computing arrangements using knowledge-based models
    • G06N5/02Knowledge representation; Symbolic representation
    • G06N5/022Knowledge engineering; Knowledge acquisition

Definitions

  • the present invention relates to the technical field of graph data processing, in particular to a graph data processing method, device, equipment, and medium.
  • the present invention provides a graph data processing method, device, equipment, and medium, so as to overcome the above-mentioned problems or at least partially solve the above-mentioned problems.
  • a graph data processing method applied to FPGA including:
  • the boundary value and weight of each of the graph data blocks are stored in the corresponding memory according to the second preset rule, so that the boundary value and the weight are used to schedule the target graph data in the graph calculation process.
  • the method before performing statistics on the target graph data according to the first preset rule, the method further includes:
  • the performing statistics on the target graph data according to a first preset rule so as to divide the target graph data into different graph data blocks and determine the boundary value and weight of each graph data block, including :
  • the target graph data is divided into different graph data blocks according to the flag value, and the boundary value and weight of each graph data block are determined.
  • performing statistics on the target graph data according to a first preset rule to determine a flag value corresponding to any data in the target graph data includes:
  • the second preset flag value is determined as the flag value corresponding to the data.
  • the process of comparing the data with the next data adjacent to the data further includes:
  • 0 is determined as the next data adjacent to the data, so as to compare the data with the next data adjacent to the data.
  • the dividing the target graph data into different graph data blocks according to the flag value and determining the boundary value and weight of each graph data block includes:
  • the current flag value is the second preset flag value
  • the current flag value is determined as the first position flag value, and the second preset flag value before the current flag value and closest to the current flag value is determined , In order to determine the value of the second location flag;
  • the number of data in the current graph data block is determined as the weight of the current graph data block.
  • the storing the boundary value and weight of each of the graph data blocks in the corresponding memory according to the second preset rule includes:
  • a graph data processing device applied to FPGA including:
  • the data acquisition module is used to acquire the target graph data to be processed
  • a data statistics module configured to perform statistics on the target graph data according to a first preset rule, so as to divide the target graph data into different graph data blocks and determine the boundary value and weight of each graph data block;
  • the data storage module is configured to store the boundary value and weight of each of the graph data blocks in the corresponding memory according to the second preset rule, so as to use the boundary value and the weight to calculate the target during the graph calculation process.
  • Graph data for scheduling is configured to store the boundary value and weight of each of the graph data blocks in the corresponding memory according to the second preset rule, so as to use the boundary value and the weight to calculate the target during the graph calculation process.
  • a graph data processing device including:
  • the memory is used to store a computer program
  • the processor is configured to execute the computer program to implement the graph data processing method disclosed above.
  • a computer-readable storage medium for storing a computer program, wherein the computer program is executed by a processor to implement the above-disclosed graph data processing method.
  • the present invention provides a graph data processing method, device, equipment, and medium. First obtain the target image data to be processed, and then perform statistics on the target image data according to the first preset rule, so as to divide the target image data into different image data blocks and determine the boundary value of each of the image data blocks And weight, and then store the boundary value and weight of each graph data block in the corresponding memory according to the second preset rule, so as to use the boundary value and the weight to compare the target graph data in the graph calculation process Schedule. It can be seen from this that the application collects statistics on the acquired target graph data to be processed according to the first preset rule, so as to divide the target graph data into different graph data blocks and determine the boundary value of each of the graph data blocks.
  • the weight and then store the boundary value and weight of the graph data block according to the second preset rule, so that the boundary value and the weight can be used to schedule the target graph data in the graph calculation process, It can quickly and accurately schedule the target graph data block, save graph data scheduling time, and improve the efficiency of graph data processing.
  • Fig. 1 is a flowchart of a graph data processing method disclosed in this application
  • FIG. 3 is a specific flow chart for determining the value and weight of the graph data disclosed in this application.
  • FIG. 4 is a schematic diagram of the structure of a graph data processing device disclosed in this application.
  • Fig. 5 is a structural diagram of a graph data processing device disclosed in this application.
  • this application proposes a graph data processing method, which can quickly and accurately schedule the target graph data block, save graph data scheduling time, and improve the efficiency of graph data processing.
  • an embodiment of the present application discloses a graph data processing method applied to FPGA, and the method includes:
  • Step S11 Obtain target image data to be processed.
  • the method further includes: arranging the target graph data in descending order; or, arranging the target graph data in descending order Arrange in order. That is, arranging the target image data in a certain order for subsequent block division, where the arranging in a certain order includes: arranging in ascending order or arranging in ascending order.
  • Step S12 Perform statistics on the target graph data according to a first preset rule, so as to divide the target graph data into different graph data blocks and determine the boundary value and weight of each graph data block.
  • performing statistics on the target graph data according to the first preset rule so as to divide the target graph data into different graph data blocks and determine the boundary value and weight of each graph data block includes: Perform statistics on the target graph data according to the first preset rule to determine the flag value corresponding to each data in the target graph data; divide the target graph data into different graph data blocks according to the flag value and determine The boundary value and weight of each of the graph data blocks.
  • first perform statistics on the target book data according to the first preset rule determine the flag value corresponding to each data in the target image data, and then divide the target image data into different images according to the flag value.
  • the data block and the boundary value and weight of each of the graph data blocks are determined.
  • Step S13 Store the boundary value and weight of each of the graph data blocks in the corresponding memory according to the second preset rule, so that the boundary value and the weight are used to perform the calculation on the target graph data in the graph calculation process. Scheduling.
  • storing the boundary value and weight of each of the graph data blocks in a corresponding memory according to a second preset rule includes: using the boundary value of the graph data block as a target Address, the boundary value and weight of each of the graph data blocks are stored in the corresponding target address, so that the boundary value and weight of each of the graph data blocks are stored in the memory. Specifically, the boundary value of each graph data block is used as the target address, and the boundary value and weight of each graph data block are stored in the corresponding target address to complete the boundary value of each graph data block. And weights are stored in memory.
  • storing the boundary value and weight of each of the graph data blocks in a corresponding memory according to a second preset rule includes: using the boundary value of the graph data block as the first A target address, the weight of each of the picture data blocks is stored in the corresponding first target address; the weight value and the boundary value are read from the first target address, and the weight value is used as The second target address is to store the boundary value and weight of each graph data block under the corresponding second target address to complete the storage of the boundary value and the weight corresponding to each graph data block into the memory. If the boundary values of some image data blocks are different but the weights are the same, the corresponding second target address is determined according to the sequence of reading time. The weight is used as the second target address for storage, so that the weight difference between adjacent image data blocks will not be too large, thereby saving image data processing time.
  • this application first obtains the target graph data to be processed, and then performs statistics on the target graph data according to the first preset rule, so as to divide the target graph data into different graph data blocks and determine each graph data
  • the boundary value and weight of each block are stored in the corresponding memory according to the second preset rule, so that the boundary value and the weight are used in the calculation process of the graph.
  • the target graph data is scheduled. It can be seen from this that this application collects statistics on the acquired target graph data to be processed according to the first preset rule, so as to divide the target graph data into different graph data blocks and determine the boundary value of each of the graph data blocks.
  • the weight and then store the boundary value and weight of the graph data block according to the second preset rule, so that the boundary value and the weight can be used to schedule the target graph data in the graph calculation process, It can quickly and accurately schedule the target graph data block, save graph data scheduling time, and improve the efficiency of graph data processing.
  • an embodiment of the present application discloses a specific graph data processing method, which is applied to FPGA, and the method includes:
  • Step S21 Obtain target image data to be processed.
  • Step S22 Perform statistics on the target graph data according to the first preset rule, and determine the flag value corresponding to each data in the target graph data.
  • the target data may be arranged in descending order or from smallest to largest to obtain the arranged target image data, and then the arranged target data may be aligned according to the first preset rule.
  • the graph data is counted, and the flag bit corresponding to each data in the target graph data is determined.
  • Performing statistics on the target map data according to the first preset rule to determine the flag value corresponding to any data in the target map data includes: comparing the data with the next data adjacent to the data, and Determine whether the data is the same as the next data adjacent to the data; if the data is the same as the next data adjacent to the data, the first preset flag value is determined as the flag value corresponding to the data; if the data is the same If the data is different from the next data adjacent to the data, the second preset flag value is determined as the flag value corresponding to the data.
  • the process of comparing the data with the next data adjacent to the data further includes: if the data is the last data of the target image data, then determining 0 as the next data adjacent to the data , In order to compare the data with the next data adjacent to the data. Specifically, the flag value corresponding to any data in the target image data is obtained by comparing the data with the next data adjacent to the data. If the data is equal to the next data adjacent to the data, Then the flag value corresponding to the data is the first preset flag value, and if the data is not equal to the next data adjacent to the data, the flag value corresponding to the data is the second preset flag value.
  • the flag value of the data processed in the current clock cycle needs to be registered first, and after the data in the next clock cycle is read, the current clock cycle The last data is compared with the first data of the next clock cycle to get the flag value of the last data of the current clock cycle.
  • Step S23 Divide the target graph data into different graph data blocks according to the flag value, and determine the boundary value and weight of each graph data block.
  • dividing the target graph data into different graph data blocks according to the flag value and determining the boundary value and weight of each graph data block includes: determining whether the current flag value is the second preset Flag value; if the current flag value is the second preset flag value, the current flag value is determined as the first position flag value, and the second preset value before the current flag value and closest to the current flag value is determined Set the flag value to determine the second position flag value; divide the target image data corresponding to the current flag value from the next flag value adjacent to the second position flag value to the current flag value of the first position flag value into one image Data block, and determine the target image data corresponding to the current flag value as the boundary value of the current image data block; determine the number of data in the current image data block as the weight of the current image data block.
  • the first preset flag value is 0, the second preset flag value is 1, and the bus bit width is 8 image data. Take the image data in two clock cycles as an example, and mark the data in each clock cycle accordingly data_0 to data_7 represent 8 pieces of data.
  • Step S24 Use the boundary value of the graph data block as a target address, and store the boundary value and weight of each graph data block in the corresponding target address, so as to store the boundary value and weight of each graph data block To the memory.
  • an embodiment of the present application discloses a graph data processing device, which is applied to an FPGA, and includes:
  • the data acquisition module 11 is used to acquire the target image data to be processed
  • the data statistics module 12 is configured to perform statistics on the target graph data according to a first preset rule, so as to divide the target graph data into different graph data blocks and determine the boundary value and weight of each graph data block;
  • the data storage module 13 is configured to store the boundary value and weight of each of the graph data blocks in the corresponding memory according to the second preset rule, so as to use the boundary value and the weight to calculate the graph in the graph calculation process.
  • the target graph data is scheduled.
  • this application first obtains the target graph data to be processed, and then performs statistics on the target graph data according to the first preset rule, so as to divide the target graph data into different graph data blocks and determine each graph data
  • the boundary value and weight of each block are stored in the corresponding memory according to the second preset rule, so that the boundary value and the weight are used in the calculation process of the graph.
  • the target graph data is scheduled. It can be seen from this that this application collects statistics on the acquired target graph data to be processed according to the first preset rule, so as to divide the target graph data into different graph data blocks and determine the boundary value of each of the graph data blocks.
  • the weight and then store the boundary value and weight of the graph data block according to the second preset rule, so that the boundary value and the weight can be used to schedule the target graph data in the graph calculation process, It can quickly and accurately schedule the target graph data block, save graph data scheduling time, and improve the efficiency of graph data processing.
  • an embodiment of the present application also discloses a graph data processing device, including: a processor 21 and a memory 22.
  • the memory 22 is used to store a computer program; the processor 21 is used to execute the computer program to implement the graph data processing method disclosed in the foregoing embodiment.
  • the embodiment of the present application also discloses a computer-readable storage medium for storing a computer program, wherein the computer program is executed by a processor to implement the following steps:
  • this application first obtains the target graph data to be processed, and then performs statistics on the target graph data according to the first preset rule, so as to divide the target graph data into different graph data blocks and determine each graph data
  • the boundary value and weight of each block are stored in the corresponding memory according to the second preset rule, so that the boundary value and the weight are used in the calculation process of the graph.
  • the target graph data is scheduled. It can be seen from this that this application collects statistics on the acquired target graph data to be processed according to the first preset rule, so as to divide the target graph data into different graph data blocks and determine the boundary value of each of the graph data blocks.
  • the weight and then store the boundary value and weight of the graph data block according to the second preset rule, so that the boundary value and the weight can be used to schedule the target graph data in the graph calculation process, It can quickly and accurately schedule the target graph data block, save graph data scheduling time, and improve the efficiency of graph data processing.
  • the following steps can be specifically implemented: arrange the target graph data in ascending order; or, arrange the The target image data is arranged in order from largest to smallest.
  • the target graph data is counted according to the first preset rule, and the target is determined The flag value corresponding to each data in the graph data; the target graph data is divided into different graph data blocks according to the mark value, and the boundary value and weight of each graph data block are determined.
  • the following steps can be specifically implemented: comparing the data with the next data adjacent to the data to determine the data Is the same as the next data adjacent to the data; if the data is the same as the next data adjacent to the data, the first preset flag value is determined as the flag value corresponding to the data; if the data and the next data are the same If the next data adjacent to the data is not the same, the second preset flag value is determined as the flag value corresponding to the data.
  • the following steps can be specifically implemented: if the data is the last data of the target graph data, then 0 is determined as The next data adjacent to the data so that the data can be compared with the next data adjacent to the data.
  • the following steps can be specifically implemented: judging whether the current flag value is the second preset flag value; if the current flag value is For the second preset flag value, the current flag value is determined as the first position flag value, and the second preset flag value before the current flag value and closest to the current flag value is determined, so as to determine the first position flag value.
  • Two position flag values the target image data corresponding to the current flag value from the next flag value adjacent to the second position flag value to the first position flag value is divided into one image data block, and the current flag value
  • the corresponding target image data is determined as the boundary value of the current image data block; the number of data in the current image data block is determined as the weight of the current image data block.
  • the following steps can be specifically implemented: taking the boundary value of the graph data block as the target address, and assigning each graph data block The boundary value and weight of is stored in the corresponding target address, so that the boundary value and weight of each of the graph data blocks are stored in the memory.
  • this application can be provided as methods, systems, or computer program products. Therefore, this application may adopt the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Moreover, this application may adopt the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program codes.
  • computer-usable storage media including but not limited to disk storage, CD-ROM, optical storage, etc.
  • These computer program instructions can also be stored in a computer-readable memory that can guide a computer or other programmable data processing equipment to work in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture including the instruction device.
  • the device implements the functions specified in one process or multiple processes in the flowchart and/or one block or multiple blocks in the block diagram.
  • These computer program instructions can also be loaded on a computer or other programmable data processing equipment, so that a series of operation steps are executed on the computer or other programmable equipment to produce computer-implemented processing, so as to execute on the computer or other programmable equipment.
  • the instructions provide steps for implementing the functions specified in one process or multiple processes in the flowchart and/or one block or multiple blocks in the block diagram.
  • the computing device includes one or more processors (CPUs), input/output interfaces, network interfaces, and memory.
  • processors CPUs
  • input/output interfaces network interfaces
  • memory volatile and non-volatile memory
  • the memory may include non-permanent memory in a computer-readable medium, random access memory (RAM) and/or non-volatile memory, such as read-only memory (ROM) or flash memory (flash RAM).
  • RAM random access memory
  • ROM read-only memory
  • flash RAM flash memory
  • Computer-readable media include permanent and non-permanent, removable and non-removable media, and information storage can be realized by any method or technology.
  • the information can be computer-readable instructions, data structures, program modules, or other data.
  • Examples of computer storage media include, but are not limited to, phase change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technology, CD-ROM, digital versatile disc (DVD) or other optical storage, Magnetic cassettes, magnetic tape magnetic disk storage or other magnetic storage devices or any other non-transmission media can be used to store information that can be accessed by computing devices. According to the definition in this article, computer-readable media does not include transitory media, such as modulated data signals and carrier waves.
  • this application can be provided as a method, a system, or a computer program product. Therefore, this application may adopt the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Moreover, this application may adopt the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program codes.
  • a computer-usable storage media including but not limited to disk storage, CD-ROM, optical storage, etc.

Landscapes

  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • Databases & Information Systems (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • General Engineering & Computer Science (AREA)
  • Software Systems (AREA)
  • Data Mining & Analysis (AREA)
  • Computer Hardware Design (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Complex Calculations (AREA)
  • Image Generation (AREA)

Abstract

一种图数据处理方法、装置、设备、介质,该方法包括:获取待处理的目标图数据;按照第一预设规则对所述目标图数据进行统计,以便将所述目标图数据分成不同的图数据块以及确定出各个所述图数据块的边界值和权重;按照第二预设规则将各个所述图数据块的边界值和权重存储到相应的内存中,以便在图计算过程中利用所述边界值和所述权重对所述目标图数据进行调度。这样在图计算过程中便可利用所述边界值和所述权重对所述目标图数据进行调度,能够快速、准确调度到目标图数据块,节约图数据调度时间,提高图数据处理的效率。

Description

一种图数据处理方法、装置、设备、介质
本申请要求于2020年02月28日提交中国专利局、申请号为202010131165.8、发明名称为“一种图数据处理方法、装置、设备、介质”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
本发明涉及图数据处理技术领域,特别涉及一种图数据处理方法、装置、设备、介质。
背景技术
在诸如万维网、社会网络、基因组分析和医学信息学等新兴应用中,图形对于表示真实的网络数据变得越来越重要。由于图数据需要存储在内存中,所以存储会受限于内存资源和图的规模,一般对于规模大的图通常需要分割成若干个子图进行存储,这样往往会出现由于图数据调度效率低而导致的图计算性能下降问题,所以在图计算中如何准确、快速地调度到目标子图块,以便完成图计算就成了一个重要问题。
发明内容
鉴于上述问题,本发明提供一种图数据处理法、装置、设备、介质,以便克服上述问题或者至少部分地解决上述问题。
一种图数据处理方法,应用于FPGA,包括:
获取待处理的目标图数据;
按照第一预设规则对所述目标图数据进行统计,以便将所述目标图数据分成不同的图数据块以及确定出各个所述图数据块的边界值和权重;
按照第二预设规则将各个所述图数据块的边界值和权重存储到相应的内存中,以便在图计算过程中利用所述边界值和所述权重对所述目标图数据进行调度。
可选地,所述按照第一预设规则对所述目标图数据进行统计之前,还包括:
将所述目标图数据按照从小到大的顺序进行排列;
或,将所述目标图数据按照从大到小的顺序进行排列。
可选地,所述按照第一预设规则对所述目标图数据进行统计,以便将所述目标图数据分成不同的图数据块以及确定出各个所述图数据块的边界值和权重,包括:
按照第一预设规则对所述目标图数据进行统计,确定出所述目标图数据中各个数据对应的标志值;
根据所述标志值将所述目标图数据分成不同的图数据块以及确定出各个所述图数据块的边界值和权重。
可选地,按照第一预设规则对所述目标图数据进行统计,确定出所述目标图数据中任一数据对应的标志值,包括:
将该数据和与该数据相邻的下一个数据进行比较,以判断该数据和与该数据相邻的下一个数据是否相同;
如果该数据和与该数据相邻的下一个数据相同,则将第一预设标志值确定为该数据对应的标志值;
如果该数据和与该数据相邻的下一个数据不相同,则将第二预设标志值确定为该数据对应的标志值。
可选地,所述将该数据和与该数据相邻的下一个数据进行比较的过程中,还包括:
如果该数据为所述目标图数据的最后一个数据,则将0确定为与该数据相邻的下一个数据,以便将该数据和与该数据相邻的下一个数据进行比较。
可选地,所述根据所述标志值将所述目标图数据分成不同的图数据块以及确定出各个所述图数据块的边界值和权重,包括:
判断当前标志值是否为所述第二预设标志值;
如果当前标志值为所述第二预设标志值,则将当前标志值确定为第一位置标志值,并确定出当前标志值之前,且与当前标志值最近的所述第二预设标志值,以便确定出第二位置标志值;
将与所述第二位置标志值相邻的下一个标志值到所述第一位置标志值 的当前标志值对应的目标图数据分为一个图数据块,并将当前标志值对应的目标图数据确定为当前图数据块的边界值;
将当前图数据块中的数据个数确定为当前图数据块的权重。
可选地,所述按照第二预设规则将各个所述图数据块的边界值和权重存储到相应的内存中,包括:
将所述图数据块的边界值作为目标地址,将各个所述图数据块的边界值和权重存入对应的目标地址下,以便将各个所述图数据块的边界值和权重存储到内存中。
一种图数据处理装置,应用于FPGA,包括:
数据获取模块,用于获取待处理的目标图数据;
数据统计模块,用于按照第一预设规则对所述目标图数据进行统计,以便将所述目标图数据分成不同的图数据块以及确定出各个所述图数据块的边界值和权重;
数据存储模块,用于按照第二预设规则将各个所述图数据块的边界值和权重存储到相应的内存中,以便在图计算过程中利用所述边界值和所述权重对所述目标图数据进行调度。
一种图数据处理设备,包括:
存储器和处理器;
其中,所述存储器,用于存储计算机程序;
所述处理器,用于执行所述计算机程序,以实现前述公开的图数据处理方法。
一种计算机可读存储介质,用于保存计算机程序,其中,所述计算机程序被处理器执行时实现前述公开的图数据处理方法。
借由上述技术方案,本发明提供的图数据处理法、装置、设备、介质。先获取待处理的目标图数据,然后按照第一预设规则对所述目标图数据进行统计,以便将所述目标图数据分成不同的图数据块以及确定出各个所述图数据块的边界值和权重,再按照第二预设规则将各个所述图数据块的边界值和权重存储到相应的内存中,以便在图计算过程中利用所述边界值和所述权重对所述目标图数据进行调度。由此可见,本申请将获取到的待处 理的目标图数据按照第一预设规则进行统计,以便将所述目标图数据分成不同的图数据块以及确定出各个所述图数据块的边界值和权重,再按照第二预设规则对所述图数据块的边界值和权重进行存储,这样在图计算过程中便可利用所述边界值和所述权重对所述目标图数据进行调度,能够快速、准确调度到目标图数据块,节约图数据调度时间,提高图数据处理的效率。
上述说明仅是本发明技术方案的概述,为了能够更清楚了解本发明的技术手段,而可依照说明书的内容予以实施,并且为了让本发明的上述和其它目的、特征和优点能够更明显易懂,以下特举本发明的具体实施方式。
附图说明
通过阅读下文优选实施方式的详细描述,各种其他的优点和益处对于本领域普通技术人员将变得清楚明了。附图仅用于示出优选实施方式的目的,而并不认为是对本发明的限制。而且在整个附图中,用相同的参考符号表示相同的部件。在附图中:
图1为本申请公开的一种图数据处理方法流程图;
图2为本申请公开的一种具体的图数据处理方法流程图;
图3为本申请公开的一种具体的图数据标志值和权重确定流程图;
图4为本申请公开的一种图数据处理装置结构示意图;
图5为本申请公开的一种图数据处理设备结构图。
具体实施方式
下面将参照附图更详细地描述本公开的示例性实施例。虽然附图中显示了本公开的示例性实施例,然而应当理解,可以以各种形式实现本公开而不应被这里阐述的实施例所限制。相反,提供这些实施例是为了能够更透彻地理解本公开,并且能够将本公开的范围完整的传达给本领域的技术人员。
目前,由于图数据需要存储在内存中,所以存储会受限于内存资源和 图的规模,一般对于规模大的图通常需要分割成若干个子图进行存储,这就面临在图计算中如何准确、快速地调度到目标子图块,以解决由图数据调度效率低而导致的图计算性能下降问题。有鉴于此,本申请提出了一种图数据处理方法,能够快速、准确调度到目标图数据块,节约图数据调度时间,提高图数据处理的效率。
参见图1所示,本申请实施例公开了一种图数据处理方法,应用于FPGA,该方法包括:
步骤S11:获取待处理的目标图数据。
在具体的实施过程中,需要先获取待处理的目标图数据,以便对所述目标图数据进行相应的后续处理。
所述按照第一预设规则对所述目标图数据进行统计之后,还包括:将所述目标图数据按照从小到大的顺序进行排列;或,将所述目标图数据按照从大到小的顺序进行排列。也即,将所述目标图数据按照一定的顺序进行排列,以便进行后续分块,其中,按照一定的顺序进行排列包括:按照由小到大的顺序排列或按照由大到小的顺序排列。
步骤S12:按照第一预设规则对所述目标图数据进行统计,以便将所述目标图数据分成不同的图数据块以及确定出各个所述图数据块的边界值和权重。
可以理解的是,在获取到所述目标图数据之后,还需要按照第一预设规则对所述目标图数据进行处理,以便将所述目标图数据分成不同的图数据块以及确定出各个所述图数据块的边界值和权重。具体的,所述按照第一预设规则对所述目标图数据进行统计,以便将所述目标图数据分成不同的图数据块以及确定出各个所述图数据块的边界值和权重,包括:按照第一预设规则对所述目标图数据进行统计,确定出所述目标图数据中各个数据对应的标志值;根据所述标志值将所述目标图数据分成不同的图数据块以及确定出各个所述图数据块的边界值和权重。也即,先按照第一预设规则对所述目标图书据进行统计,确定出所述目标图数据中各个数据对应的标志值,再根据所述标志值将所述目标图数据分成不同的图数据块以及确定出各个所述图数据块的边界值和权重。
步骤S13:按照第二预设规则将各个所述图数据块的边界值和权重存储到相应的内存中,以便在图计算过程中利用所述边界值和所述权重对所述目标图数据进行调度。
在得到各个图数据块的边界值和权重之后,还需要按照第二预设规则将各个所述图数据块的边界值和权重存储到相应的内存中,以便在图计算过程中利用所述边界值和所述权重对所述目标图数据进行调度。
在第一种具体的实施方式中,所述按照第二预设规则将各个所述图数据块的边界值和权重存储到相应的内存中,包括:将所述图数据块的边界值作为目标地址,将各个所述图数据块的边界值和权重存入对应的目标地址下,以便将各个所述图数据块的边界值和权重存储到内存中。具体的,就是将各个所述图数据块的边界值作为目标地址,将各个所述图数据块的边界值和权重存入对应的目标地址下,以完成将各个所述图数据块的边界值和权重存储到内存中。
在第二种具体的实施方式中,所述按照第二预设规则将各个所述图数据块的边界值和权重存储到相应的内存中,包括:将所述图数据块的边界值作为第一目标地址,将各个所述图数据块的权重存入对应的第一目标地址下;从所述第一目标地址下读出所述权重值和所述边界值,并将所述权重值作为第二目标地址,将各个图数据块的边界值和权重存储到相应的第二目标地址下,完成将各个所述图数据块对应的所述边界值和所述权重存入到内存中。若有的图数据块的边界值不相同,但是权重相同,则按照读取时间的先后顺序确定对应的第二目标地址。以权重作为第二目标地址进行存储,使得相邻图数据块之间的权重差值不会太大,从而节约图数据处理时间。
可见,本申请先获取待处理的目标图数据,然后按照第一预设规则对所述目标图数据进行统计,以便将所述目标图数据分成不同的图数据块以及确定出各个所述图数据块的边界值和权重,再按照第二预设规则将各个所述图数据块的边界值和权重存储到相应的内存中,以便在图计算过程中利用所述边界值和所述权重对所述目标图数据进行调度。由此可见,本申请将获取到的待处理的目标图数据按照第一预设规则进行统计,以便将所 述目标图数据分成不同的图数据块以及确定出各个所述图数据块的边界值和权重,再按照第二预设规则对所述图数据块的边界值和权重进行存储,这样在图计算过程中便可利用所述边界值和所述权重对所述目标图数据进行调度,能够快速、准确调度到目标图数据块,节约图数据调度时间,提高图数据处理的效率。
参见图2所示,本申请实施例公开了一种具体的图数据处理方法,应用于FPGA,该方法包括:
步骤S21:获取待处理的目标图数据。
步骤S22:按照第一预设规则对所述目标图数据进行统计,确定出所述目标图数据中各个数据对应的标志值。
在具体的实施过程中,可以先将所述目标数据由大到小或由小到大的顺序进行排列,得到排列后的目标图数据,然后按照第一预设规则对排列后的所述目标图数据进行统计,确定出所述目标图数据中各个数据对应的标志位。
按照第一预设规则对所述目标图数据进行统计,确定出所述目标图数据中任一数据对应的标志值,包括:将该数据和与该数据相邻的下一个数据进行比较,以判断该数据和与该数据相邻的下一个数据是否相同;如果该数据和与该数据相邻的下一个数据相同,则将第一预设标志值确定为该数据对应的标志值;如果该数据和与该数据相邻的下一个数据不相同,则将第二预设标志值确定为该数据对应的标志值。将该数据和与该数据相邻的下一个数据进行比较的过程中,还包括:如果该数据为所述目标图数据的最后一个数据,则将0确定为与该数据相邻的下一个数据,以便将该数据和与该数据相邻的下一个数据进行比较。具体的,所述目标图数据中任一数据对应的标志值是通过将该数据与和该数据相邻的下一个数据比较得到的,如果该数据与和该数据相邻的下一个数据相等,则该数据对应的标志值为第一预设标志值,如果该数据与和该数据相邻的下一个数据不相等,则该数据对应的标志值为第二预设标志值。由于FPGA中一个时钟周期内能处理的数据量和总线位宽相关,所以当前时钟周期内处理的数据的标志 值需要先进行寄存,等下一个时钟周期的数据读取之后,将当前时钟周期的最后一个数据与下一个时钟周期的第一个数据比较才能得到当前时钟周期的最后一个数据的标志值。
步骤S23:根据所述标志值将所述目标图数据分成不同的图数据块以及确定出各个所述图数据块的边界值和权重。
在确定出所述目标图数据中各个数据的标志值之后,需要根据所述标志值将所述目标图数据分成不同的图数据块,并确定出各个所述图数据块的边界值和权重。其中,所述根据所述标志值将所述目标图数据分成不同的图数据块以及确定出各个所述图数据块的边界值和权重,包括:判断当前标志值是否为所述第二预设标志值;如果当前标志值为所述第二预设标志值,则将当前标志值确定为第一位置标志值,并确定出当前标志值之前,且与当前标志值最近的所述第二预设标志值,以便确定出第二位置标志值;将与所述第二位置标志值相邻的下一个标志值到所述第一位置标志值的当前标志值对应的目标图数据分为一个图数据块,并将当前标志值对应的目标图数据确定为当前图数据块的边界值;将当前图数据块中的数据个数确定为当前图数据块的权重。参见图3所示,为图数据标志值和权重确定流程图。第一预设标志值为0,第二预设标志值为1,总线位宽为8个图数据,以两个时钟周期内的图数据为例,且将各个时钟周期内的数据相应的记号data_0到data_7表示8个数据。将时钟周期1中的data_0与data_1比较,得到data_0=data_1=1,则data_0的标志值flag[0]=0,以此类推,得到flag[1]=0,flag[2]=0,data_3=1,data_4=3,data_3与data_4不相等,则flag[3]=1,以此类推,得到其余数据的标志值,时钟周期1的data_7与时钟周期2的data_0比较,得到时钟周期1中的flag[7]=0,假设时钟周期2的data_7是所述目标图数据的最后一个数据,则时钟周期2的data_7不等于0,则时钟周期2中的flag[7]=1,根据以上标志值,将时钟周期1中的data_0到data_3的4个数据分为一个图数据块,相应的权重为4,边界值为1;将时钟周期1中的data_4到时钟周期2中的data_0的5个数据分为一个图数据块,相应的权重为5,边界值为3;将时钟周期2中的data_1的1个数据分为一个图数据块,相应的权重为1,边界值为10;将时钟周 期2中的data_2到data_4的3个数据分为一个图数据块,相应的权重为3,边界值为23;将时钟周期2中的data_5的1个数据分为一个图数据块,相应的权重为1,边界值为59;将时钟周期2中的data_6到data_7的2个数据分为一个图数据块,相应的权重为2,边界值为60。
步骤S24:将所述图数据块的边界值作为目标地址,将各个所述图数据块的边界值和权重存入对应的目标地址下,以便将各个所述图数据块的边界值和权重存储到内存中。
参见图4所示,本申请实施例公开了一种图数据处理装置,应用于FPGA,包括:
数据获取模块11,用于获取待处理的目标图数据;
数据统计模块12,用于按照第一预设规则对所述目标图数据进行统计,以便将所述目标图数据分成不同的图数据块以及确定出各个所述图数据块的边界值和权重;
数据存储模块13,用于按照第二预设规则将各个所述图数据块的边界值和权重存储到相应的内存中,以便在图计算过程中利用所述边界值和所述权重对所述目标图数据进行调度。
可见,本申请先获取待处理的目标图数据,然后按照第一预设规则对所述目标图数据进行统计,以便将所述目标图数据分成不同的图数据块以及确定出各个所述图数据块的边界值和权重,再按照第二预设规则将各个所述图数据块的边界值和权重存储到相应的内存中,以便在图计算过程中利用所述边界值和所述权重对所述目标图数据进行调度。由此可见,本申请将获取到的待处理的目标图数据按照第一预设规则进行统计,以便将所述目标图数据分成不同的图数据块以及确定出各个所述图数据块的边界值和权重,再按照第二预设规则对所述图数据块的边界值和权重进行存储,这样在图计算过程中便可利用所述边界值和所述权重对所述目标图数据进行调度,能够快速、准确调度到目标图数据块,节约图数据调度时间,提高图数据处理的效率。
进一步的,参见图5所示,本申请实施例还公开了一种图数据处理设备,包括:处理器21和存储器22。
其中,所述存储器22,用于存储计算机程序;所述处理器21,用于执行所述计算机程序,以实现前述实施例中公开的图数据处理方法。
其中,关于上述图数据处理方法的具体过程可以参考前述实施例中公开的相应内容,在此不再进行赘述。
进一步的,本申请实施例还公开了一种计算机可读存储介质,用于保存计算机程序,其中,所述计算机程序被处理器执行时实现以下步骤:
获取待处理的目标图数据;按照第一预设规则对所述目标图数据进行统计,以便将所述目标图数据分成不同的图数据块以及确定出各个所述图数据块的边界值和权重;按照第二预设规则将各个所述图数据块的边界值和权重存储到相应的内存中,以便在图计算过程中利用所述边界值和所述权重对所述目标图数据进行调度。
可见,本申请先获取待处理的目标图数据,然后按照第一预设规则对所述目标图数据进行统计,以便将所述目标图数据分成不同的图数据块以及确定出各个所述图数据块的边界值和权重,再按照第二预设规则将各个所述图数据块的边界值和权重存储到相应的内存中,以便在图计算过程中利用所述边界值和所述权重对所述目标图数据进行调度。由此可见,本申请将获取到的待处理的目标图数据按照第一预设规则进行统计,以便将所述目标图数据分成不同的图数据块以及确定出各个所述图数据块的边界值和权重,再按照第二预设规则对所述图数据块的边界值和权重进行存储,这样在图计算过程中便可利用所述边界值和所述权重对所述目标图数据进行调度,能够快速、准确调度到目标图数据块,节约图数据调度时间,提高图数据处理的效率。
本实施例中,所述计算机可读存储介质中保存的计算机子程序被处理器执行时,可以具体实现以下步骤:将所述目标图数据按照从小到大的顺序进行排列;或,将所述目标图数据按照从大到小的顺序进行排列。
本实施例中,所述计算机可读存储介质中保存的计算机子程序被处理 器执行时,可以具体实现以下步骤:按照第一预设规则对所述目标图数据进行统计,确定出所述目标图数据中各个数据对应的标志值;根据所述标志值将所述目标图数据分成不同的图数据块以及确定出各个所述图数据块的边界值和权重。
本实施例中,所述计算机可读存储介质中保存的计算机子程序被处理器执行时,可以具体实现以下步骤:将该数据和与该数据相邻的下一个数据进行比较,以判断该数据和与该数据相邻的下一个数据是否相同;如果该数据和与该数据相邻的下一个数据相同,则将第一预设标志值确定为该数据对应的标志值;如果该数据和与该数据相邻的下一个数据不相同,则将第二预设标志值确定为该数据对应的标志值。
本实施例中,所述计算机可读存储介质中保存的计算机子程序被处理器执行时,可以具体实现以下步骤:如果该数据为所述目标图数据的最后一个数据,则将0确定为与该数据相邻的下一个数据,以便将该数据和与该数据相邻的下一个数据进行比较。
本实施例中,所述计算机可读存储介质中保存的计算机子程序被处理器执行时,可以具体实现以下步骤:判断当前标志值是否为所述第二预设标志值;如果当前标志值为所述第二预设标志值,则将当前标志值确定为第一位置标志值,并确定出当前标志值之前,且与当前标志值最近的所述第二预设标志值,以便确定出第二位置标志值;将与所述第二位置标志值相邻的下一个标志值到所述第一位置标志值的当前标志值对应的目标图数据分为一个图数据块,并将当前标志值对应的目标图数据确定为当前图数据块的边界值;将当前图数据块中的数据个数确定为当前图数据块的权重。
本实施例中,所述计算机可读存储介质中保存的计算机子程序被处理器执行时,可以具体实现以下步骤:将所述图数据块的边界值作为目标地址,将各个所述图数据块的边界值和权重存入对应的目标地址下,以便将各个所述图数据块的边界值和权重存储到内存中。
本领域内的技术人员应明白,本申请的实施例可提供为方法、***、或计算机程序产品。因此,本申请可采用完全硬件实施例、完全软件实施 例、或结合软件和硬件方面的实施例的形式。而且,本申请可采用在一个或多个其中包含有计算机可用程序代码的计算机可用存储介质(包括但不限于磁盘存储器、CD-ROM、光学存储器等)上实施的计算机程序产品的形式。
本申请是参照根据本申请实施例的方法、设备(***)、和计算机程序产品的流程图和/或方框图来描述的。应理解可由计算机程序指令实现流程图和/或方框图中的每一流程和/或方框、以及流程图和/或方框图中的流程和/或方框的结合。可提供这些计算机程序指令到通用计算机、专用计算机、嵌入式处理机或其他可编程数据处理设备的处理器以产生一个机器,使得通过计算机或其他可编程数据处理设备的处理器执行的指令产生用于实现在流程图一个流程或多个流程和/或方框图一个方框或多个方框中指定的功能的装置。
这些计算机程序指令也可存储在能引导计算机或其他可编程数据处理设备以特定方式工作的计算机可读存储器中,使得存储在该计算机可读存储器中的指令产生包括指令装置的制造品,该指令装置实现在流程图一个流程或多个流程和/或方框图一个方框或多个方框中指定的功能。
这些计算机程序指令也可装载到计算机或其他可编程数据处理设备上,使得在计算机或其他可编程设备上执行一系列操作步骤以产生计算机实现的处理,从而在计算机或其他可编程设备上执行的指令提供用于实现在流程图一个流程或多个流程和/或方框图一个方框或多个方框中指定的功能的步骤。
在一个典型的配置中,计算设备包括一个或多个处理器(CPU)、输入/输出接口、网络接口和内存。
存储器可能包括计算机可读介质中的非永久性存储器,随机存取存储器(RAM)和/或非易失性内存等形式,如只读存储器(ROM)或闪存(flash RAM)。存储器是计算机可读介质的示例。
计算机可读介质包括永久性和非永久性、可移动和非可移动媒体可以由任何方法或技术来实现信息存储。信息可以是计算机可读指令、数据结构、程序的模块或其他数据。计算机的存储介质的例子包括,但不限于相 变内存(PRAM)、静态随机存取存储器(SRAM)、动态随机存取存储器(DRAM)、其他类型的随机存取存储器(RAM)、只读存储器(ROM)、电可擦除可编程只读存储器(EEPROM)、快闪记忆体或其他内存技术、只读光盘只读存储器(CD-ROM)、数字多功能光盘(DVD)或其他光学存储、磁盒式磁带,磁带磁磁盘存储或其他磁性存储设备或任何其他非传输介质,可用于存储可以被计算设备访问的信息。按照本文中的界定,计算机可读介质不包括暂存电脑可读媒体(transitory media),如调制的数据信号和载波。
还需要说明的是,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,从而使得包括一系列要素的过程、方法、商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、商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,并不排除在包括要素的过程、方法、商品或者设备中还存在另外的相同要素。
本领域技术人员应明白,本申请的实施例可提供为方法、***或计算机程序产品。因此,本申请可采用完全硬件实施例、完全软件实施例或结合软件和硬件方面的实施例的形式。而且,本申请可采用在一个或多个其中包含有计算机可用程序代码的计算机可用存储介质(包括但不限于磁盘存储器、CD-ROM、光学存储器等)上实施的计算机程序产品的形式。
以上仅为本申请的实施例而已,并不用于限制本申请。对于本领域技术人员来说,本申请可以有各种更改和变化。凡在本申请的精神和原理之内所作的任何修改、等同替换、改进等,均应包含在本申请的权利要求范围之内。

Claims (10)

  1. 一种图数据处理方法,其特征在于,应用于FPGA,包括:
    获取待处理的目标图数据;
    按照第一预设规则对所述目标图数据进行统计,以便将所述目标图数据分成不同的图数据块以及确定出各个所述图数据块的边界值和权重;
    按照第二预设规则将各个所述图数据块的边界值和权重存储到相应的内存中,以便在图计算过程中利用所述边界值和所述权重对所述目标图数据进行调度。
  2. 根据权利要求1所述的图数据处理方法,其特征在于,所述按照第一预设规则对所述目标图数据进行统计之前,还包括:
    将所述目标图数据按照从小到大的顺序进行排列;
    或,将所述目标图数据按照从大到小的顺序进行排列。
  3. 根据权利要求1所述的图数据处理方法,其特征在于,所述按照第一预设规则对所述目标图数据进行统计,以便将所述目标图数据分成不同的图数据块以及确定出各个所述图数据块的边界值和权重,包括:
    按照第一预设规则对所述目标图数据进行统计,确定出所述目标图数据中各个数据对应的标志值;
    根据所述标志值将所述目标图数据分成不同的图数据块以及确定出各个所述图数据块的边界值和权重。
  4. 根据权利要求3所述的图数据处理方法,其特征在于,按照第一预设规则对所述目标图数据进行统计,确定出所述目标图数据中任一数据对应的标志值,包括:
    将该数据和与该数据相邻的下一个数据进行比较,以判断该数据和与该数据相邻的下一个数据是否相同;
    如果该数据和与该数据相邻的下一个数据相同,则将第一预设标志值确定为该数据对应的标志值;
    如果该数据和与该数据相邻的下一个数据不相同,则将第二预设标志值确定为该数据对应的标志值。
  5. 根据权利要求4所述的图数据处理方法,其特征在于,所述将该数 据和与该数据相邻的下一个数据进行比较的过程中,还包括:
    如果该数据为所述目标图数据的最后一个数据,则将0确定为与该数据相邻的下一个数据,以便将该数据和与该数据相邻的下一个数据进行比较。
  6. 根据权利要求4所述的图数据处理方法,其特征在于,所述根据所述标志值将所述目标图数据分成不同的图数据块以及确定出各个所述图数据块的边界值和权重,包括:
    判断当前标志值是否为所述第二预设标志值;
    如果当前标志值为所述第二预设标志值,则将当前标志值确定为第一位置标志值,并确定出当前标志值之前,且与当前标志值最近的所述第二预设标志值,以便确定出第二位置标志值;
    将与所述第二位置标志值相邻的下一个标志值到所述第一位置标志值的当前标志值对应的目标图数据分为一个图数据块,并将当前标志值对应的目标图数据确定为当前图数据块的边界值;
    将当前图数据块中的数据个数确定为当前图数据块的权重。
  7. 根据权利要求1所述的图数据处理方法,其特征在于,所述按照第二预设规则将各个所述图数据块的边界值和权重存储到相应的内存中,包括:
    将所述图数据块的边界值作为目标地址,将各个所述图数据块的边界值和权重存入对应的目标地址下,以便将各个所述图数据块的边界值和权重存储到内存中。
  8. 一种图数据处理装置,其特征在于,应用于FPGA,包括:
    数据获取模块,用于获取待处理的目标图数据;
    数据统计模块,用于按照第一预设规则对所述目标图数据进行统计,以便将所述目标图数据分成不同的图数据块以及确定出各个所述图数据块的边界值和权重;
    数据存储模块,用于按照第二预设规则将各个所述图数据块的边界值和权重存储到相应的内存中,以便在图计算过程中利用所述边界值和所述权重对所述目标图数据进行调度。
  9. 一种图数据处理设备,其特征在于,包括:
    存储器和处理器;
    其中,所述存储器,用于存储计算机程序;
    所述处理器,用于执行所述计算机程序,以实现权利要求1至7任一项所述的图数据处理方法。
  10. 一种计算机可读存储介质,其特征在于,用于保存计算机程序,其中,所述计算机程序被处理器执行时实现如权利要求1至7任一项所述的图数据处理方法。
PCT/CN2020/126349 2020-02-28 2020-11-04 一种图数据处理方法、装置、设备、介质 WO2021169386A1 (zh)

Priority Applications (1)

Application Number Priority Date Filing Date Title
US17/797,876 US20230334094A1 (en) 2020-02-28 2020-11-04 Graph data processing method, apparatus, and device, and medium

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN202010131165.8 2020-02-28
CN202010131165.8A CN111400521B (zh) 2020-02-28 2020-02-28 一种图数据处理方法、装置、设备、介质

Publications (1)

Publication Number Publication Date
WO2021169386A1 true WO2021169386A1 (zh) 2021-09-02

Family

ID=71430441

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2020/126349 WO2021169386A1 (zh) 2020-02-28 2020-11-04 一种图数据处理方法、装置、设备、介质

Country Status (3)

Country Link
US (1) US20230334094A1 (zh)
CN (1) CN111400521B (zh)
WO (1) WO2021169386A1 (zh)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN114282073A (zh) * 2022-03-02 2022-04-05 支付宝(杭州)信息技术有限公司 数据存储方法及装置、数据读取方法及装置

Families Citing this family (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN111400521B (zh) * 2020-02-28 2022-06-07 苏州浪潮智能科技有限公司 一种图数据处理方法、装置、设备、介质
US20240037150A1 (en) * 2022-08-01 2024-02-01 Qualcomm Incorporated Scheduling optimization in sequence space

Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN103970879A (zh) * 2014-05-16 2014-08-06 中国人民解放军国防科学技术大学 一种调整数据块存储位置的方法及***
CN109492131A (zh) * 2018-09-18 2019-03-19 华为技术有限公司 一种图数据存储方法及装置
WO2019127300A1 (zh) * 2017-12-25 2019-07-04 深圳云天励飞技术有限公司 数据存储方法、电子设备及存储介质
CN110688055A (zh) * 2018-07-04 2020-01-14 清华大学 一种大图计算中数据访问方法及***
CN111400521A (zh) * 2020-02-28 2020-07-10 苏州浪潮智能科技有限公司 一种图数据处理方法、装置、设备、介质

Family Cites Families (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2016195455A1 (ko) * 2015-06-04 2016-12-08 엘지전자(주) 그래프 기반 변환을 이용하여 비디오 신호를 처리하는 방법 및 장치
CN108681482B (zh) * 2018-05-16 2020-12-01 腾讯科技(深圳)有限公司 基于图数据的任务调度方法和装置
CN109508413A (zh) * 2018-10-22 2019-03-22 中国银行股份有限公司 作业调度展示方法及装置、存储介质及电子设备
CN110188099A (zh) * 2019-05-17 2019-08-30 深圳前海微众银行股份有限公司 一种数据管理方法及装置

Patent Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN103970879A (zh) * 2014-05-16 2014-08-06 中国人民解放军国防科学技术大学 一种调整数据块存储位置的方法及***
WO2019127300A1 (zh) * 2017-12-25 2019-07-04 深圳云天励飞技术有限公司 数据存储方法、电子设备及存储介质
CN110688055A (zh) * 2018-07-04 2020-01-14 清华大学 一种大图计算中数据访问方法及***
CN109492131A (zh) * 2018-09-18 2019-03-19 华为技术有限公司 一种图数据存储方法及装置
CN111400521A (zh) * 2020-02-28 2020-07-10 苏州浪潮智能科技有限公司 一种图数据处理方法、装置、设备、介质

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN114282073A (zh) * 2022-03-02 2022-04-05 支付宝(杭州)信息技术有限公司 数据存储方法及装置、数据读取方法及装置

Also Published As

Publication number Publication date
US20230334094A1 (en) 2023-10-19
CN111400521A (zh) 2020-07-10
CN111400521B (zh) 2022-06-07

Similar Documents

Publication Publication Date Title
WO2021169386A1 (zh) 一种图数据处理方法、装置、设备、介质
US10311230B2 (en) Anomaly detection in distributed ledger systems
JP6991983B2 (ja) 機械学習システムをトレーニングする方法及びシステム
CN105022670A (zh) 一种云计算平台中的异构分布式任务处理***及其处理方法
JP2013541063A (ja) 単語のユーザー挙動数の予測
CN106569892B (zh) 资源调度方法与设备
CN113516247A (zh) 参数校准方法、量子芯片的控制方法、装置及***
CN106648839B (zh) 数据处理的方法和装置
CN113516246A (zh) 参数优化方法、量子芯片的控制方法及装置
CN107391230B (zh) 一种确定虚拟机负载的实现方法和装置
CN109033365A (zh) 一种数据处理方法及相关设备
CN111832693A (zh) 神经网络层运算、模型训练方法、装置及设备
CN105867886A (zh) 一种写表格的方法及装置
CN111125157B (zh) 查询数据的处理方法、装置、存储介质及处理器
CN115935723A (zh) 用于实现氮化镓制备场景下的设备组合分析方法及***
CN110909085A (zh) 数据处理方法、装置、设备及存储介质
JP2024050143A (ja) 量子シミュレーション実行時のノード割当プログラム、量子シミュレーション実行時のノード割当方法、および情報処理装置
CN109492086A (zh) 一种答案输出方法、装置、电子设备及存储介质
WO2021147382A1 (zh) 一种检测任务执行方法及设备
CN110019497A (zh) 一种数据读取方法及装置
CN108228560A (zh) 一种数据类型的确定方法及装置
CN109582295B (zh) 一种数据处理方法、装置、存储介质及处理器
CN108427692A (zh) 周序列号的计算方法及装置
CN106649059B (zh) 数据统计方法及装置
CN111913805B (zh) 一种cpu利用率计算方法及装置

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: 20922439

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: 20922439

Country of ref document: EP

Kind code of ref document: A1

122 Ep: pct application non-entry in european phase

Ref document number: 20922439

Country of ref document: EP

Kind code of ref document: A1