Xyce: Open-Source Circuit Simulation
| Presenter: |
Eric Keiter |
Sandia National Laboratories |
| Start: |
09:45 (30 minutes) |
Abstract
This talk provides an overview of the open-source analog simulation tool, Xyce, which was designed from the ground-up to perform large-scale circuit analysis. Current capabilities of the simulation tool will be discussed, including the analysis methods, device models, and parallel implementation. Other topics will include recent improvements in compatibility with modern process design kits (PDKs). After this broad overview, the second half of the talk will describe a new Xyce feature, a Python-based model interface. This new interface, called Xyce-PyMi, allows the user to develop models in Python and then run them directly in Xyce. This has enabled a lot of our recent research in data-driven model, which are often based on Machine Learning (ML) techniques. Most ML-based research is done in Python, and uses popular python libraries such as TensorFlow and PyTorch. With this new interface, ML-based models can be run directly in Xyce.
Biography
This talk provides an overview of the open-source analog simulation tool, Xyce, which was designed from the ground-up to perform large-scale circuit analysis. Current capabilities of the simulation tool will be discussed, including the analysis methods, device models, and parallel implementation. Other topics will include recent improvements in compatibility with modern process design kits (PDKs). After this broad overview, the second half of the talk will describe a new Xyce feature, a Python-based model interface. This new interface, called Xyce-PyMi, allows the user to develop models in Python and then run them directly in Xyce. This has enabled a lot of our recent research in data-driven model, which are often based on Machine Learning (ML) techniques. Most ML-based research is done in Python, and uses popular python libraries such as TensorFlow and PyTorch. With this new interface, ML-based models can be run directly in Xyce.
