placement optimization
1 article · search the full text for this term
-
Optimizing Routing and Placement of VLSI Circuits with Differential Algorithms and Neural Networks
Abstract: The performance of modern VLSI systems is heavily influenced by power constraints, necessitating precise power estimation and effective optimization techniques. Traditional methods, such as gate-level simulations, are often slow and computationally intensive. This paper introduces DRPENN (Differential Algorithm for Routing and Placement Optimization using Neural Networks), an innovative solution that combines a Switching Activity Estimator (SAE) with a neural network-assisted differential algorithm. By leveraging toggle rates from simulations to train …
Published in Journal of VLSI Design Tools and Technology · Vol. 14, Issue 2, 2024 · pp. 14–20 Read article