Abstract:As a typical artificial intelligence algorithm, the convolutional neural network (CNN) is widely used in the Internet of Things (IoT) system. In order to improve the computing ability of an IoT CPU, this paper designs a reconfigurable CNN-accelerated coprocessor based on the RISC-V instruction set. The interconnection structure of the acceleration chain designed by the predecessors is optimized, and the accelerator is connected to the RISC-V CPU core in the form of a coprocessor. The corresponding instruction of the coprocessor is designed and the instruction compiling environment is established. Through the inline assembly in the C language, the coprocessor instructions are called, coprocessor acceleration library functions are established, and common algorithms in the IoT system are implemented on the coprocessor. Finally, resource consumption evaluation and performance analysis of the coprocessor are completed on a Xilinx FPGA. The evaluation results show that the reconfigurable CNN-accelerated coprocessor only consumes 8534 LUTS, accounting for 47.6% of the total SoC system. The number of instruction cycles required to implement functions such as convolution and pooling based on the designed coprocessor instructions is better than using the standard instruction set, and the acceleration ratio of convolution is 6.27 times that of the standard instruction set.

Custom Embedded Processor for Access Network Protocols Based on Reduced Instruction Set Computer

Design of a Media-Enhanced and Reconfigurable RISC ,Microprocessor

The Implementation of Custom RISC Core for Access Network Protocols

Reconfigurable Designs for Networking Silicon.

Intelligence-networked Automatic Instrumentation Based on an Embedded

Design of a Reconfigurable Network Interface Processor

Design and Implementation of Multi-processor Network Access Controller for Network-on-Chip

UIC: A Unified and Scalable Chip Integrating Neuromorphic Computation and General Purpose Processor

Test Time Reduction Reusing Multiple Processors in a Network-on-Chip Based Architecture

VLSI Implementation of a 32bit Embedded Reduced-Instruction-set-computer

Design and VLSI Implementation of an Application Specific Instruction Set Security Processor

A 101.4 GOPS/W Reconfigurable and Scalable Control-Centric Embedded Processor for Domain-Specific Applications

Design of a Custom Instruction Set Processor and Its Tool Chain

Design Of Multithreaded Coprocessor Ip Core For Embedded Soc Chip

A 61 Μa/mhz Reconfigurable Application-Specific Processor and System-on-chip for Internet-of-Things

Design and Simulation of a RISC-Based 32-bit Embedded On-Board Computer

A Context-Based Simulation Tool for Design and Evaluation of Network Processors

Efficient custom instructions generation for system-level design

A hierarchical reconfigurable micro-coded multi-core processor for IoT applications

The Design Of A Rapid Prototyping Platform For Hardware/Software Co-Design

A Reconfigurable Convolutional Neural Network-Accelerated Coprocessor Based on RISC-V Instruction Set