Acceleration system and design method of clustering algorithm based on fpga

Abstract

The invention discloses an FPGA-based clustering algorithm acceleration system and a design method thereof. The method comprises the steps of obtaining a key code of each algorithm through a profiling technology; detailing the key code of each algorithm and extracting same function logic (a common operator); redesigning a code structure by using a blocking technology to increase the utilization rate of data locality and reduce the off-chip access frequency; designing an extended semantic instruction set, realizing function logic parts corresponding to the instruction set, and finishing a key code function through operations of fetching, decoding and execution of instructions; designing an acceleration framework of an accelerator and generating an IP core; and transplanting an operation system to a development board, and finishing cooperative work of software and hardware in the operation system. Various clustering algorithms can be supported and the flexibility and universality of a hardware accelerator can be improved; and the code of each algorithm is reconstructed by adopting the blocking technology to reduce the off-chip access frequency so as to reduce the influence of the off-chip access bandwidth on the acceleration effect of the accelerator.

Description

technical field [0001] The invention relates to an algorithm hardware accelerator, in particular to an FPGA-based clustering algorithm acceleration system with good versatility and high flexibility and a design method thereof. Background technique [0002] Clustering algorithm is a kind of unsupervised machine learning algorithm, which is widely used in various fields, such as: market research, pattern recognition, data mining, image processing, customer segmentation, web document classification, etc. According to the different division methods, commonly used clustering algorithms can be divided into the following categories: division methods, hierarchical methods, density-based methods, grid-based methods, and model-based methods. In different application fields and when dealing with different types of data, different clustering algorithms need to be used for cluster analysis to obtain better clustering results. For example, for spherically distributed data, the K-Means al...

AI Technical Summary

Problems solved by technology

Current hardware acceleration solutions rarely consider the impact of off-chip memory access on the acceleration effect. In reality, off-chip memory access is inevitable and the speed of off-chip memory access is very slow, making bandwidth an important factor restricting algorithm acceleration.

Images