Login / Signup

Novel data dependent divider circuit block implementation for complex division and area critical applications.

Udayan S PatankarMiguel E FloresAnts Koel
Published in: Scientific reports (2023)
This article elaborates on the state-of-the-art novel Udayan S. Patankar (USP)-Awadhoot algorithm for distinctive implementation area improvement for area-critical electronic applications. The proposed USP-Awadhoot divider is a digit recurrence class, but it can be flexibly implemented as a restoring or nonrestoring algorithm. The implementation example indicates the use of the Baudhayan-Pythagoras triplet method in association with the proposed USP-Awadhoot divider. The triplet method provides an easy way to generate Mat_Term1, Mat_Term2, and T_Term, which are further utilized with the proposed USP-Awadhoot divider. The USP-Awadhoot divider is implemented in three parts. First is preprocessing circuit stage for executing a dynamic separate scaling operation on input operands, ensuring the inputs are in the correct form. Second is the processing circuit stage for implementing the conversion logic expressed by the Awadhoot matrix, and third is the postprocessing circuit stage for recombining the individual results into the final result. The proposed divider works upto 285 MHz frequency with a power estimation of 3.366 W, also significantly improves the chip area requirements over those of the commercially and noncommercially implemented solutions.
Keyphrases
  • primary care
  • quality improvement
  • preterm infants
  • healthcare
  • machine learning
  • gestational age
  • deep learning
  • high throughput
  • big data
  • energy transfer
  • data analysis