Rounding method for indivisible floating point division radication

A rounding and division technology, applied in the direction of instruments, electrical digital data processing, digital data processing components, etc., can solve the problems of low hardware cost, long sequence, high hardware cost, etc., and achieve easy hardware implementation and debugging, algorithm Clear and easy to understand, low hardware cost effect

Active Publication Date: 2010-02-17
C SKY MICROSYST CO LTD
View PDF0 Cites 10 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0011] In order to overcome the shortcomings of existing microprocessors, such as high logic complexity, high hardware cost, and long sequence when dealing with indivisible results in division/extractio...

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
  • Rounding method for indivisible floating point division radication
  • Rounding method for indivisible floating point division radication
  • Rounding method for indivisible floating point division radication

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0036] refer to Figure 1 ~ Figure 2 , a rounding method for the square root of non-divisible floating-point division, comprising the following steps:

[0037] 1) The floating-point number input to the floating-point coprocessor performs SRT division / square root algorithm, and the floating-point number performs a loop operation to obtain the quotient / root result;

[0038] 2) In the quotient / root result obtained in step 1), determine the effective number of the floating-point number according to the position where the first 1 occurs;

[0039] 3) Subtract the result of the quotient / root result in step 1) from the result of the significant number and the remainder obtained by the SRT division / square root algorithm as the rounding judgment information; subtract the quotient / root obtained in step 1) from the result obtained in step 2) The result of significant number is left-shifted, and the number obtained is recorded as the remaining digit, and the position of the decimal point ...

Embodiment 2

[0126] This embodiment is aimed at the floating-point single-precision square root operation.

[0127] 1) Perform radix-4SRT division algorithm for two single-precision floating-point numbers input to the floating-point coprocessor, and obtain a 28-bit root result after 14 loop operations: 0.1a 25 a 24 ...a 4 0100 (binary representation), the remainder is greater than 0.

[0128] 2) For the root result obtained in step 1), determine the 24-bit effective number of the single-precision floating-point number according to the position where the first 1 appears: 1a 25 a 24 ...a 4 0 (binary representation). The result of subtracting the effective number from the root in step 1) is: 0.0...00100 (binary representation), this number is used as rounding judgment information.

[0129] 3) The root obtained in step 1) minus the effective number obtained in step 2) is left-shifted, and the position of the decimal point is set after the lowest digit of the effective number indicated in...

Embodiment 3

[0134] The floating-point number of the present embodiment is a double-precision floating-point number, and in the step 1), 28 loop operations are performed to obtain 56 quotient / root results; in step 2), the effective number of the double-precision floating-point number starts from the first 1 A total of 53-bit binary numbers are intercepted, and the intercepted effective number is used as the benchmark number of the rounding operation carried out in step 4) to step 7); in step 3), the double-precision floating-point number is obtained according to the position where the first 1 appears. Two (0.xx) or one (0.x) remaining bits are used for rounding judgment.

[0135] Other steps of this embodiment are the same as in Embodiment 3.

[0136] This embodiment is for floating-point double-precision division operations.

[0137] 1) Perform base 4SRT division algorithm for a double-precision floating-point number input to the floating-point coprocessor, and obtain a 56-bit quotient r...

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 relates to a rounding method for indivisible floating point division radication. The rounding method comprises the following steps: (1) executing an SRT division/radication algorithm toa floating point number input to a floating point coprocessor, and executing a loop operation to the floating point number and then obtaining a quotient/root result; (2) determining the effective number of the floating point number according to the appearance position of the first 1; (3) reducing the effective number result and a remainder obtained by the SRT division/radication algorithm from thequotient/root result of the step (1), and taking the difference as rounding judging information; and executing the left shift operation to the difference obtained by reducing the effective number result of the step (2) from the quotient/root of the step (1), and then taking the obtained number as a residue place; (4)-(7) selecting four kinds of different rounding modes to round; (8) if the rounding operation executed in step (4) to (7) causes the quotient/root result to have a carry bit or a borrow bit, modifying the floating point exponent part, wherein the final rounding result needs to execute the mantissa normalization. The rounding method can reduce the logic complexity, reduce the same hardware cost, and shorten the time sequence.

Description

technical field [0001] The invention relates to a microprocessor architecture design, in particular to a floating point operation unit. Background technique [0002] With the widespread application of embedded microprocessors in the fields of communication and multimedia, the requirements for their floating-point computing capabilities are getting higher and higher, and the speed and complexity of the usual software floating-point computing execution methods are far from meeting the actual needs. The floating-point coprocessor, which mainly performs floating-point operations by hardware, came into being. Floating-point division square root operation is an important part of floating-point calculation. The floating-point division square root implemented by hardware is 50~100 times faster than the software implementation method, which has great advantages. [0003] Because the floating-point division square root algorithm is complex and the sequence is long, the redundant rema...

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): G06F7/535
Inventor 刘磊葛海通孟建熠严晓浪
Owner C SKY MICROSYST 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