High-cardinal-number approximate Booth encoding method and mixed-cardinal-number Booth encoding approximate multiplier

Abstract

The invention discloses a high-cardinal-number approximate Booth encoding method and a mixed-cardinal-number Booth encoding approximate multiplier, and belongs to the technical field of integrated circuits. According to the high-cardinal-number approximate Booth encoding method, m bits with lower weights in n-bit multiplicand are encoded, two incomplete partial products A and B are obtained in combination with the n-bit multiplicand, and an approximate encoding result is obtained after addition. The mixed-radix Booth coding approximate multiplier is combined with a precise Booth coding moduleand a high-radix approximate Booth coding module to respectively obtain a precise partial product and an approximate partial product; a partial product array is formed by combining sign extension bitsgenerated by a sign bit extension algorithm provided by the invention, the final calculation result of the approximate multiplier is obtained by compressing and adding the partial product array, andin addition, an error model of the approximate multiplier is deduced to obtain a precision index. The structural complexity of the multiplier is reduced while high calculation precision is guaranteedthrough high-cardinal-number approximate Booth encoding, and hardware design is simplified due to the fact that symbol extension bits avoid accumulation of a large number of identical numbers.

Description

technical field [0001] The invention belongs to the technical field of integrated circuits, and relates to a high-radix approximate Booth coding method and a mixed-radix Booth code approximate multiplier based on the high-radix approximate Booth coding method. Background technique [0002] With the rise of new technologies such as big data, cloud computing, and the Internet of Things, computer systems are increasingly being used to interact with the physical world. Although semiconductor technology and low-power design techniques are also developing, the total energy consumption of computer systems is still growing at an alarming rate in order to process the ever-increasing amount of information. Nowadays, power consumption and energy consumption have become important factors that have to be considered in chip design. For high-performance computing equipment, such as servers, high-performance processors, etc., excessive power consumption will cause serious heat dissipation ...

AI Technical Summary

Problems solved by technology

Traditional accurate Booth coded multipliers usually use low-radix Booth codes to generate partial products. Although the coding unit is relatively simple, the number of coding units required is large, and the number of partial products generated is large, which is not conducive to subsequent partial products. Compression module design

At the same time, high-radix Booth encoding can significantly reduce the number of partial products, but the encoding circuit and partial product generation circuit are more complicated

Images