16 * 8 approximate multiplier on-chip dynamic computing system based on Booth coding

Abstract

The invention discloses a 16 * 8 approximate multiplier on-chip dynamic calculation system based on Booth coding, and belongs to the technical field of calculation, reckoning or counting. The method comprises the following steps: reading effective bit width of input operation data through a data preloading module; inputting the read effective bit width into a corresponding data calculation mode dynamic pre-analysis module, dividing the input data into different types according to the effective bit width, and taking the input data as an important basis for configuring an approximation scheme; inputting the previously obtained data bit width into an allowable error calculation module, and calculating a maximum allowable error; enabling an approximate multiplier structure dynamic configuration operation module to give a basic approximation scheme according to the data type; comparing the error corresponding to the scheme with the maximum allowable error, continuously adjusting the approximation scheme according to the comparison result, finally obtaining the most suitable scheme, and calculating the approximation result. And the most suitable approximate configuration scheme is found in an error tolerance range, and multiplication is dynamically configured in real time, so that the selection of precision and power consumption is more flexible.

Description

technical field [0001] The invention relates to a dynamically configurable approximate multiplication computing technology, in particular to a Booth coding-based 16*8 approximate multiplier on-chip dynamic computing system, which belongs to the technical field of calculation, calculation or counting. Background technique [0002] In recent years, people pay more and more attention to energy-saving design. This is because a large number of applications require low-power custom designs, but the amount of data that these hardware needs to process has increased dramatically. In order to meet these two requirements at the same time, approximate computing has become A solution that works. [0003] With the increasing popularity of applications related to human perception, such as neural networks and machine learning, the amount of calculation of on-chip hardware is gradually increasing, but the accuracy does not require complete accuracy and correctness. This is because neural net...

AI Technical Summary

Problems solved by technology

[0006] The object of the invention of the present invention is to aim at the deficiency of above-mentioned background technology, provide a kind of 16*8 approximation multiplier on-chip dynamic calculation system based on Booth coding, in approximate calculation technology, adjust approximate multiplication dynamically according to the data effective bit width of input system The configuration of the operation, find the most suitable approximate configuration scheme within the tolerance range of the error, realize the dynamic configuration of the approximate multiplication operation, make the selection of precision and power consumption more flexible, and solve the poor precision of the existing approximate multiplication dynamic configuration scheme question

Images