Hardware Implementation of High Precision Floating Point Transcend Function

DOI

10.2991/tlicsc-18.2018.6How to use a DOI?

Keywords

Floating point, transcendental function, CORDIC algorithm, pipeline architecture.

Abstract

In this paper, the combination of CORDIC algorithm and mathematical method is used to complete the implementation of 128-bit floating-point exponential function and logarithmic function. The implementation of exponential functions and logarithmic functions is involved in speech recognition, image processing, and radar detection, but the computational domain of [-6, 6] is completed in current hardware implementations. It is too narrow in practice. Here we use the CORDIC algorithm combined with the mathematical method to expand. In this design, the calculation domain of the logarithm function is 128 bits of all floating-point numbers and the calculation domain of the exponential function is [-8192, 8192]. The practicality of hardware implementation transcendental functions is improved by extending the computational domain. In the overall hardware architecture, based on the Xilinx ISE14.7 development environment, the Virtex7 development board, the combined maximum frequency can reach 353.189MHz.

Copyright

© 2018, the Authors. Published by Atlantis Press.

Open Access

This is an open access article distributed under the CC BY-NC license (http://creativecommons.org/licenses/by-nc/4.0/).

Download article (PDF)

TY  - CONF
AU  - Jun Zhang
AU  - Mingjiang Wang
AU  - Nanxin Hou
PY  - 2018/12
DA  - 2018/12
TI  - Hardware Implementation of High Precision Floating Point Transcend Function
BT  - Proceedings of the 2018 International Conference on Transportation & Logistics, Information & Communication, Smart City (TLICSC 2018)
PB  - Atlantis Press
SP  - 34
EP  - 39
SN  - 1951-6851
UR  - https://doi.org/10.2991/tlicsc-18.2018.6
DO  - 10.2991/tlicsc-18.2018.6
ID  - Zhang2018/12
ER  -