Deep-learning cell-delay modeling for static timing analysis

doi:10.1016/j.asej.2022.101828

Deep-learning cell-delay modeling for static timing analysis

2023-02-01

4

OA

AI

DOI：10.1016/j.asej.2022.101828

Original

Save

View PDF

Abstract

En 中文

Delay and transition timetables plus voltage waveforms are used to characterize standard cell delays. More accurate models explode cell library size and degrades design flow performance. Our proposed deep learning non-linear delay model, DL-NLDM, technique outperformed 7x7 NLDM-LUT in average percentage errors with up to 1.4% error compared to SPICE and outperformed the non-standard 100x100 NLDMLUT in maximum percentage errors. The proposed DL Autoencoder-based waveform compression outperformed singular value decomposition by 1.79x. Additionally, a novel DL waveform-delay model, DLWFDM, models cell delays using encoded waveforms instead of delay and transition time. DL-WFDM outperformed DL-NLDM in maximum delay percentage errors.

Keywords:

Electronic Design Automation

Machine Learning

Deep Learning

Autoencoders

Cell Characterization

Effective Current Source Model

Deep-learning cell-delay modeling for static timing analysis

Abstract

AI Summary

Journal

Organization