FPGA-based space-time diagram neural network accelerator structure

Abstract

The invention discloses a space-time diagram neural network accelerator structure based on an FPGA, a tensor and vector double acceleration module parallel processing mode is designed, a vector acceleration module executes sparse matrix multiplication or element-by-element multiplication and addition operation, and a tensor acceleration module executes dense matrix multiplication, bias term addition and operation of different activation functions. The system control module controls the tensor acceleration module and the vector acceleration module to complete calculation corresponding to a calculation time sequence according to the set calculation time sequence, and the data flow control module controls the data flow direction between the tensor acceleration module and the vector acceleration module so as to complete calculation in a circulating mode. According to the design of the whole architecture, efficient processing of a three-dimensional space-time diagram neural network is achieved, and calculation of a multifunctional function is achieved in a simple mode; the acceleration calculation of multiple neural network modes can be realized, so that the method can be downwards compatible with the acceleration of multiple network models such as a graph convolutional neural network GCN and a gated recursive unit GRU, and has higher universality.

Description

technical field [0001] The invention relates to the field of spatiotemporal graph data processing, in particular to an FPGA-based spatiotemporal graph neural network accelerator structure. Background technique [0002] A spatiotemporal graph is a three-dimensional network data structure with time and space, which is a kind of dynamic graph. It is widely present in graph data containing temporal information such as social networks and sensor networks. Unlike static graphs where the information of each node and connection does not change over time, dynamic graphs have an additional time dimension, and its nodes and connections may change over time. The spatiotemporal graph has important application and research value in recommendation systems, traffic forecasting, and epidemic spread forecasting. With the continuous growth of the amount of data, artificial intelligence (AI) algorithms have achieved great development and are widely used in many fields such as image, voice, an...

AI Technical Summary

Problems solved by technology

Therefore, the existing technology generally accelerates a certain neural network structure, and it is difficult to adapt to network models containing multiple computing styles.

The space-time graph neural network is a model that includes both the graph convolutional neural network GCN and the recurrent neural network GRU. Its calculation functions include dense matrix operations, sparse matrix operations, element-by-element multiplication and addition, and integration of various activation functions. Its complex computational characteristics make it difficult to achieve good performance on accelerators of existing architectures

Images