Multi-core heterogeneous CPU - CPU - FPGA architecture

Abstract

The invention discloses a multi-core heterogeneous CPU - CPU - FPGA architecture. An architecture is characterized of having a separate parallel CPU and parallel GPU, having their own SDRAM, being capable of accessing each other's SDRAM, and they are in the outside of the FPGA structure; a GPU is connected to a chip set via an I / O bus and then connected to the CPU via an I / O bridge. The CPU consists of an ALU, a register file, a flash cache, and a bus interface.

Description

technical field [0001] The invention relates to a multi-core heterogeneous CPU-GPU-FPGA system architecture. Background technique [0002] In order to achieve the main purpose of the research on vehicle safety assisted driving technology, it is necessary to grasp the status of the vehicle, the road, and the surrounding vehicles through various sensors (passive infrared camera, industrial camera, microwave radar, laser radar, etc.) installed on the vehicle. , to assist the driver to increase the ability to perceive the environment and provide early warning signals. [0003] Fusion of information from multiple sensors can achieve performance that is difficult to obtain from a single sensor, mainly reflected in the improvement of information redundancy and complementarity. The performance advantages of information fusion are mainly reflected in: improving the credibility of information and the detectability of targets, expanding the search range of time and space, reducing the...

AI Technical Summary

Problems solved by technology

[0006] Although dedicated computing units such as GPUs have low operating frequency, more cores and parallel computing capabilities, the overall performance / chip area ratio and performance / power consumption ratio are high, but they are far from being fully utilized.

Images