Dynamic Maximal Clique Enumeration with FPGA-HBM Pipelining

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing incremental maximal clique enumeration methods for dynamic graphs lack support for pipelined computation, failing to leverage the performance advantages of FPGA hardware resources effectively, especially when dealing with large-scale, dynamically changing graph data.

Innovation Solution

A dynamic maximal clique enumeration device and method utilizing an FPGA with HBM, incorporating a matrix computing unit, sequence computing unit, and update computing unit, which perform pipelined-type incremental maximal clique computation by updating a graph adjacency matrix, reconstructing candidate cliques, and executing update tasks in parallel using FIFOs and BRAM blocks.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Productivity

If existing incremental maximal clique enumeration methods are used, then the algorithm can handle dynamic graph changes, but the computing efficiency and speed are insufficient for large-scale graph data

Engineering Contradiction:
Improvecomputing efficiencyVSAvoidcomputation time
Core Design Contradiction:
ProductivityVSLoss of time

Solution Approach 1:

The patent segments the incremental maximal clique enumeration process into three distinct pipeline stages: (1) edge change processing that updates the adjacency matrix, (2) candidate clique reconstruction that identifies affected cliques, and (3) maximal clique update that computes the final result. Each stage operates independently and can be processed in parallel, enabling efficient utilization of FPGA hardware resources and significantly improving computing efficiency for large-scale graph data.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The patent implements a dynamic pipeline architecture where data flows continuously through the three computation stages. The system dynamically processes edge additions and deletions in real-time, maintaining up-to-date maximal clique information without requiring full recomputation. This dynamic approach allows the system to adapt to changing graph structures while maintaining high throughput and low latency.

Inventive Principle:
Principle #15Dynamics

2Productivity

If traditional computing architectures are used, then the implementation is simpler, but the performance advantages of FPGA hardware resources are not fully utilized

Engineering Contradiction:
Improvecomputing speedVSAvoidhardware architecture complexity
Core Design Contradiction:
ProductivityVSDevice complexity

Solution Approach 1:

The patent replaces traditional software-based incremental maximal clique enumeration with a hardware-accelerated FPGA implementation. The computation logic is directly mapped to FPGA logic elements, allowing parallel execution of multiple graph operations simultaneously. This substitution of mechanical/computational approach enables full utilization of FPGA's parallel processing capabilities, achieving significant speedup despite increased hardware complexity.

Inventive Principle:
Principle #28Mechanics substitution (Replace mechanical system)

Solution Approach 2:

The patent introduces an intermediate adjacency matrix representation that serves as a bridge between the input edge changes and the output maximal clique results. This intermediate structure allows the FPGA to efficiently process edge additions and deletions by maintaining updated neighbor information, enabling fast computation without requiring complex direct mapping from edge changes to clique updates.

Inventive Principle:
Principle #24Intermediary (Mediator)

3Reliability

If incremental computation is implemented, then the system can respond to graph changes in real-time, but the lack of pipelined processing reduces overall task efficiency

Engineering Contradiction:
Improvereal-time responseVSAvoidtask processing efficiency
Core Design Contradiction:
ReliabilityVSProductivity

Solution Approach 1:

The patent implements a dynamic pipeline architecture where data flows continuously through the three computation stages: edge change processing, candidate clique reconstruction, and maximal clique update. Each stage operates independently and can be processed in parallel, enabling efficient utilization of FPGA hardware resources. This pipelined dynamic processing maintains real-time responsiveness while significantly improving overall task processing efficiency through concurrent execution.

Inventive Principle:
Principle #15Dynamics

Data Source

PatentUS12380176B2Dynamic maximal clique enumeration device and method based on FPGA with HBM
Publication Date: 2025.08.05 ZHEJIANG LAB
  • US12380176B2 patent drawing
  • US12380176B2 patent drawing
  • US12380176B2 patent drawing

AI summary

Disclosed in the present invention are a dynamic maximal clique enumeration device and method based on an FPGA with an HBM, the method including: the HBM stores a dynamic edge flow, a complete graph adjacency matrix, and candidate cliques; a matrix computing unit updates the complete graph adjacency matrix based on the dynamic edge flow, transmits the updated complete graph adjacency matrix to the HBM for storage, and determines header nodes, of which the corresponding candidate clique needs to be updated; a sequence computing unit constructs, according to the updated complete graph adjacency matrix and each header node to be updated, the sorted data set for reconstructing candidate cliques by data block sequencing; and an update computing unit executes, in parallel, an update task of the candidate clique corresponding to each header node to be updated based on the sorted data set, transmits the updated candidate cliques to the HBM for storage, and transmits the updated candidate cliques to the PC host to extract maximal cliques by means of a filtering operation. The present invention supports the computation of pipelined-type incremental maximal clique, thereby improving the overall computing efficiency of the task.