Machine Learning/AI for EDA

This page lists resources related to all forms of Machine Learning/AI algorithms and frameworks developed for various Electronic Design and Automation (EDA) problems. Feel free to browse through.

Code/framework packages may be legally protected. Kindly go through the licensing details included in the package. Contact diraclab@ece.iitr.ac.in for any issues/access.

NN-assisted Compact Models

  • Coming soon...

    FALCON (Fast Augmented learning COmpact model with Neural networks) is a class of NN-assisted compact models that utilise a hybrid of physics equations and neural networks to provide superior runtime performance with similar accuracy as athat of traditional compact models, while retaining parametric tuning flexibility and complete compatibility with existing workflows and tools. This version is meant for all common-multi-gate devices like FinFETs, Gate-All-Around FETs, Forksheet FETs etc.

  • Model code is available for use. Please write to diraclab@ece.iitr.ac.in requesting access.

    FALCON (Fast Augmented learning COmpact model with Neural networks) is a class of NN-assisted compact models that utilise a hybrid of physics equations and neural networks to provide superior runtime performance with similar accuracy as athat of traditional compact models, while retaining parametric tuning flexibility and complete compatibility with existing workflows and tools. FALCON-Q utilizes a core that puts increased emphasis on correct internal charge predictions as compared to the regular FALCON. This version is meant for all common-multi-gate devices like FinFETs, Gate-All-Around FETs, Forksheet FETs etc.

Contact Us

Reach out to DiRac Lab

Location:

W-202/1,
Department of Electronics and Communication Engineering,
Indian Institute of Technology Roorkee