CORDIC Hardware Acceleration Using DMA-Based ISA Extension

Erez Manor, Avrech Ben-David, Shlomo Greenberg

Research output: Contribution to journalArticlepeer-review

7 Scopus citations

Abstract

The use of RISC-based embedded processors aimed at low cost and low power is becoming an increasingly popular ecosystem for both hardware and software development. High-performance yet low-power embedded processors may be attained via the use of hardware acceleration and Instruction Set Architecture (ISA) extension. Recent publications of AI have demonstrated the use of Coordinate Rotation Digital Computer (CORDIC) as a dedicated low-power solution for solving nonlinear equations applied to Neural Networks (NN). This paper proposes ISA extension to support floating-point CORDIC, providing efficient hardware acceleration for mathematical functions. A new DMA-based ISA extension approach integrated with a pipeline CORDIC accelerator is proposed. The CORDIC ISA extension is directly interfaced with a standard processor data path, allowing efficient implementation of new trigonometric ALU-based custom instructions. The proposed DMA-based CORDIC accelerator can also be used to perform repeated array calculations, offering a significant speedup over software implementations. The proposed accelerator is evaluated on Intel Cyclone-IV FPGA as an extension to Nios processor. Experimental results show a significant speedup of over three orders of magnitude compared with software implementation, while applied to trigonometric arrays, and outperforms the existing commercial CORDIC hardware accelerator.

Original languageEnglish
Article number4
JournalJournal of Low Power Electronics and Applications
Volume12
Issue number1
DOIs
StatePublished - 1 Mar 2022

Keywords

  • CORDIC
  • FPGA
  • Hardware accelerator
  • ISA-Extension
  • Low-power
  • Neural networks
  • RISC

ASJC Scopus subject areas

  • Electrical and Electronic Engineering

Fingerprint

Dive into the research topics of 'CORDIC Hardware Acceleration Using DMA-Based ISA Extension'. Together they form a unique fingerprint.

Cite this