Graph structure-aware graph processing method and device
What is AI technical title?
AI technical title is built by PatSnap AI team. It summarizes the technical point description of the patent document.
A technology of graph processing and graph structure, applied in the field of data processing, to achieve the effect of reducing the average number of updates, reducing loading, and reducing I/O overhead
Active Publication Date: 2020-10-30
HUAZHONG UNIV OF SCI & TECH
View PDF3 Cites 0 Cited by
Summary
Abstract
Description
Claims
Application Information
AI Technical Summary
This helps you quickly interpret patents by identifying the three key elements:
Problems solved by technology
Method used
Benefits of technology
Problems solved by technology
Existing systems such as PowerSwitch, PowerLyra, GraphH, etc. have not solved these problems
Method used
the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View more
Image
Smart Image Click on the blue labels to locate them in the text.
Viewing Examples
Smart Image
Click on the blue label to locate the original text in one second.
Reading with bidirectional positioning of images and text.
Smart Image
Examples
Experimental program
Comparison scheme
Effect test
Embodiment 1
[0036] Such as figure 1 As shown, the present invention provides a graph structure-aware graph processing method, including the following steps:
[0037] S1: Divide hot blocks and cold blocks based on the heat value of the vertex;
[0038] S2: Obtain the block state degree of hot block and cold block respectively based on the iterative calculation results of two adjacent vertices;
[0039] S3: in progress every T 1 In the case of round iterations, the hot blocks are re-divided. If there are still hot blocks after the division, the cold blocks and / or hot blocks are adaptively scheduled for calculation. After the division, there are no hot blocks and the entire graph does not converge Next, every T 2 round iteration;
[0040] S4: Determine the convergence of the entire graph after iterative calculation based on the sum of all block state degrees.
[0041] For ease of understanding, each step of the graph processing method of the present invention will be described in detail...
Embodiment 2
[0064] Embodiment 2 is a further optimization and improvement of Embodiment 1, and the repeated content will not be repeated.
[0065] Before iteratively processing the graph vertices, the graph vertices are preprocessed. Such as image 3 As shown, a method for preprocessing graph vertices includes the following steps:
[0066] S1: Identify and judge all vertices whose degrees are zero, mark them as dead vertices and store them in the dead block sequentially.
[0067] S2: Graph vertex initialization: Create two graph vertex degree tables and two graph vertex value tables. Among them, the two graph vertex degree tables are vertex table and edge table respectively. The point table is used to store the ID, in-degree, out-degree and thermal value of graph vertices. The edge table is used to store graph vertex IDs and neighbor graph vertex IDs. The vertex value table in the first graph is used to store the calculation result of the vertex this time, and the vertex value table ...
Embodiment 3
[0078] This embodiment is a further improvement on Embodiment 1 and Embodiment 2, and repeated content will not be repeated here.
[0079] A graph structure-aware graph processing method, including a graph data preprocessing stage, a graph structure dynamic division stage, and a graph processing adaptive scheduling stage. Among them, in the graph data preprocessing stage, the vertex heat is calculated according to the vertex in-degree and out-degree, the vertex heat is read sequentially, and the corresponding vertices are divided into hot blocks and cold blocks. The dynamic partition stage of the graph structure, each T 1 The round iteration calls the graph partition function to re-partition the hot blocks. In the case of no hot blocks after redivision, judge whether the entire graph is converged. When the entire graph converges, the calculation output results are ended. Only cold block calculations are scheduled when the entire graph has not converged. Preferably, when th...
the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More
PUM
Login to View More
Abstract
The invention relates to a graph processing method for graph structure sensing, which comprises the steps of dividing a hot block and a cold block based on a heat value of a vertex; obtaining block state degrees of the hot block and the cold block based on a calculation result of the vertex; performing re-division on the hot blocks in each T1 turn of iteration, selecting a block with the highest state degree in the hot blocks for calculation, and selecting a block with the highest state degree in the cold blocks and the hot blocks in each T2 turn of iteration for calculation; and judging the convergence of the graph based on the sum of the state degrees. The graph processing device for graph structure sensing comprises a central processor and a graph processor which are mutually interacted, wherein the graph processor performs descending sorting on the vertexes according to the heat values and / or vertex state degrees, and the central processor divides the vertexes to the hot blocks andthe cold blocks based on the heat values, calculates the block state degrees of the heat blocks and the cold blocks, then performs iteration and judges the convergence of the graph after the iterative calculation. The method can reduce the cache missing rate and I / O resource overhead and accelerate the vertex convergence. The device can reduce the iterative calculation time, and does not generateadditional communication overhead at the same time.
Description
technical field [0001] The present invention relates to the technical field of data processing, in particular to a graph structure-aware graph processing method and graph processing device. Background technique [0002] Graph is a topological relational model that abstracts big data or human behavior to build a computing model, which is composed of nodes (ie, vertices) and relationships (ie, edges formed by vertex connections) as basic elements. The essence of a graph is a data structure composed of multiple data that are interrelated. This data structure can effectively express the correlation between data, and by obtaining the correlation of data, useful information can be extracted from massive data with a lot of noise. For example, for a consumer's original purchase behavior, there are two types of nodes: user and product, and the relationship refers to the purchase behavior. The purchasing behavior of countless consumers forms a big data structure with correlation. B...
Claims
the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More
Application Information
Patent Timeline
Application Date:The date an application was filed.
Publication Date:The date a patent or application was officially published.
First Publication Date:The earliest publication date of a patent with the same application number.
Issue Date:Publication date of the patent grant document.
PCT Entry Date:The Entry date of PCT National Phase.
Estimated Expiry Date:The statutory expiry date of a patent right according to the Patent Law, and it is the longest term of protection that the patent right can achieve without the termination of the patent right due to other reasons(Term extension factor has been taken into account ).
Invalid Date:Actual expiry date is based on effective date or publication date of legal transaction data of invalid patent.
Login to View More 
Login to View More