Design method of double-precision floating-point divider and divider

A design method and double-precision technology, applied in the field of dividers, can solve the problem that the performance advantages of double-precision floating-point dividers cannot be exerted, and achieve the effect of floor area overhead and high performance

Active Publication Date: 2014-05-21
PHYTIUM TECH CO LTD
View PDF1 Cites 21 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

Although this method can achieve complete pipelining, due to the less use of division operations, the perf

Method used

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View more

Image

Smart Image Click on the blue labels to locate them in the text.
Viewing Examples
Smart Image
  • Design method of double-precision floating-point divider and divider
  • Design method of double-precision floating-point divider and divider
  • Design method of double-precision floating-point divider and divider

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0071] A design method of a double-precision floating-point number divider and the divider of the present invention will be described in detail below in conjunction with the embodiments and the accompanying drawings.

[0072] The design method and divider of a kind of double-precision floating-point number divider invented, adopt the method of one look-up table + two Goldschmidt iterations to design the double-precision floating-point divider, wherein the lookup table is also obtained by using the quadratic polynomial approximation algorithm, But its area is much smaller than the area of ​​the lookup table used by the original method. After the lookup table, a seed value Rf with a smaller precision than the original method will be obtained. In the subsequent Goldschmidt, two iterations are used to obtain the same value as the original The method is the same precision, but both iterations use the same hardware unit (2 multipliers of 60x60). In general, the method we adopt has a...

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to view more

PUM

No PUM Login to view more

Abstract

The invention provides a design method of a double-precision floating-point divider and a divider. The method includes the steps that (1) a seed value Rf of a reciprocal function 1/X is calculated in a min-max quadratic polynomial approach algorithm, wherein X is a 53-bit magnitude portion of a divisor; (2) on the basis of twice Goldschmidt iteration of a hardware multiplexing method, an ultimate precise result is obtained. The divider comprises a first portion used for calculating the seed value Rf of the reciprocal function 1/X and a second portion connected with output of the first portion and used for calculating the ultimate precise result. The divider is high in performance and low in area overhead, by means of the hardware multiplexing method, under the condition that performance is seldom reduced, large area advantages are obtained, and the design method and the divider are especially suitable for being used in the field of embedment-type micro processors strict at area requirement.

Description

technical field [0001] The present invention relates to a divider. In particular, it relates to a design method of a high-performance, low-area overhead double-precision floating-point number divider and the divider. Background technique [0002] The division algorithm is mainly divided into two types: one is a slow algorithm, and the other is a fast algorithm. The slow algorithm is a method based on digital iteration, each iteration generates a bit, and the convergence speed is linear convergence. Restoring remainder method, non-restoring remainder method and SRT method are all slow algorithms. The fast algorithm usually uses a look-up table method to obtain a seed value close to the quotient value, and multiplication is used as the basic iterative step. Each iteration generates multiple quotients, and the convergence speed is logarithmic convergence. Both Newton-Raphson and Goldschmidt are fast algorithms. [0003] According to the definition of double-precision floati...

Claims

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to view more

Application Information

Patent Timeline
no application Login to view more
IPC IPC(8): G06F7/535G06F17/50
Inventor 郭炜崔鲁平李光赫魏继增
Owner PHYTIUM TECH CO LTD
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Try Eureka
PatSnap group products

Browse by: Latest US Patents, China's latest patents, Technical Efficacy Thesaurus, Application Domain, Technology Topic.

© 2024 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap