Share Your Research, Maximize Your Social Impacts
High Performance MAC based on Multi-Level Approximate Compressors with Balanced Error Accumulation
Journal: IMPACT : International Journal of Research in Engineering & Technology ( IMPACT : IJRET ) (Vol.10, No. 6)Publication Date: 2022-06-30
Authors : Aneesath. K P Manivannan. R Arun Prasath. N; G. Ranjithkumar;
Page : 1-8
Keywords : Approximate Computing; Convolutional Neural Network; Image Processing; Low-Power Circuit Design; Multiplier;
Source : Download
Find it from : Google Scholar
Abstract
It is presented a unique approximate computing approach for implementing energy-efficient multiply-accumulate (MAC) processing. We use different approximate multipliers in an interleaved manner to mitigate mistakes in the opposite direction during accumulate operations, in contrast to previous efforts that suffer from error accumulation restricting the approximation range. We first build approximation 4-2 compressors
Other Latest Articles
- VISCERAL MOBILIZATION AS MANAGEMENT OPTION FOR LOW BACK PAIN - A SYSTEMATIC REVIEW
- Star–in–Coloring of Theta and Plus Graphs
- ROLE OF REHAB PROFESSIONALS IN COVID-19 CRISES
- A NATIONAL NEED: REHABILITATION SPECIALIST IN SUBSTANCE ABUSE DISORDER
- Road Safety System Based on Arduino Built with Eye Blink Sensor
Last modified: 2022-09-02 19:07:17