Full-flow 128-bit-accuracy floating-point accumulator based on full expansion

An accumulator and floating-point technology, which is applied to calculations using the number system and calculations using non-contact manufacturing equipment, and can solve problems such as undiscovered floating-point accumulators

Inactive Publication Date: 2010-10-13
NAT UNIV OF DEFENSE TECH
View PDF2 Cites 15 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0009] In the existing accumulator design, there is no report of using the FPGA chip to realize the fully expanded, fully pipelined 128-bit precision floating point accumulator

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
  • Full-flow 128-bit-accuracy floating-point accumulator based on full expansion
  • Full-flow 128-bit-accuracy floating-point accumulator based on full expansion
  • Full-flow 128-bit-accuracy floating-point accumulator based on full expansion

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0053] figure 1 is a general block diagram of an existing accumulator. The existing floating-point accumulator is designed according to the idea of ​​cyclically performing floating-point addition operations, and is mainly composed of an order unit, an adder and a normalization unit. The normalization section consists of a leading zero count section, a normalization shift section, and a normalization rounding section. The pair-order component receives operand A and operand B input from the outside, and shifts the mantissa of the operand with a small exponent to the right according to the exponent difference between the two operands, and aligns it with the mantissa of the operand with a large exponent. The adder reads two operands from the pair-order part to add, and inputs the addition result to the leading zero counting part and the normalized shifting part; the leading zero counting part calculates the number of leading zeros according to the addition result, and puts the l...

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 discloses a full-flow 128-bit-accuracy floating-point accumulator based on full expansion. The main technical problem to be solved is to provide an accumulator for realizing lossless accuracy accumulation. The full-flow 128-bit-accuracy floating-point accumulator based on full expansion comprises a normalized module and a full-expansion floating-point accumulator module which comprises a mantissa addition module, a memory module and a rapid carry module, the memory module comprises four memories, a counter and four selectors, the mantissa addition module comprises a complementation part, a shifting part, an operator A register, an operator B register, an addition module, an addition result register and a first flag upgrading part, and the rapid carry module comprises a carry termination factor address generating part, a carry termination factor address register, a carry termination factor generating part, a carry information register, a carry termination factor register, a carry adder, a carry addition result register and a second flag upgrading part. The accumulator of the invention has improved accuracy and speed.

Description

technical field [0001] The invention relates to a microprocessor in the field of integrated circuits, in particular to a fully expanded and fully pipelined 128-bit precision floating-point accumulator. Background technique [0002] Since the birth of the computer, its computing speed and computing power have been greatly improved, from 1 billion times / second (Gigascale) of Cray2 in 1985 to the current petascale / second (Petascale). It is estimated that by 2015, it will reach Exascale, which means that the computer will be able to perform 1018 standard IEEE754 double-precision floating-point operations per second. The improvement of the computer's computing speed and computing power means that it can meet the needs of practical applications in a wider range, but this also brings a new problem: the floating-point operation process will produce rounding errors, and with the calculation scale The error will continue to accumulate, which may cause the final calculation result to ...

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
Patent Type & Authority Applications(China)
IPC IPC(8): G06F7/50
Inventor 窦勇雷元武郭松
Owner NAT UNIV OF DEFENSE TECH
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