Fast methods simulate nonlinear circuits for advanced energy computing
Fast simulation of nonlinear deep resistive networks for energy-based computation
Emerging Technologies
Summary
Simulating nonlinear electronic circuits is usually slow because of complex device behavior. The authors improved simulation methods to work with more realistic circuit parts like real diodes, making simulations much faster but still accurate. Their approach works for large circuit networks and helps train energy-based analog computing systems efficiently. This advancement opens the door to designing and training big analog computer circuits that use actual components.
What this means in practice
- •For analog circuit designers: Simulate and train large scalable analog circuits with realistic nonlinear devices much faster than traditional methods.
- •For machine learning engineers: Train energy-based analog neural networks efficiently using realistic device models to speed up tasks like image recognition.
Authors
Filip Osana, Julie Grollier, Damien Querlioz
Abstract
Deep resistive networks are electronic energy-based systems in which computation is performed by the steady-state voltages of nonlinear circuits. Nonlinear devices enable expressive input-output transformations, but make the circuit equilibria costly to compute during simulation and training. Recent coordinate-descent solvers have achieved large speedups over SPICE-class circuit simulators, but only for nonlinearities modeled as ideal-diode models. This mathematical simplification is not sufficient for practical analog circuits. Here we extend coordinate-descent simulation to realistic monotone nonlinearities, including Shockley diodes, antiparallel diode pairs, and piecewise-linear current-voltage characteristics. With neighboring voltages fixed, each node update remains a scalar Kirchhoff-law solve. Single-exponential characteristics admit closed-form Lambert-(W) updates, while more general monotone characteristics can be handled with scalar root-finding methods. Across networks with one to three hidden layers and hidden widths from 64 to 1024, the solver reproduces matched SPICE steady-state voltages with relative $L_1$ errors below $1.1\times10^{-4}$ for 90% of validation samples, while achieving speedups up to $(1.7\times10^3)$. We further train a $1568\times100\times20$ double-Shockley network on MNIST, reaching about 3.0% test error and reducing the per-epoch training time by roughly $4.4\times10^2$. Together, these results establish a practical route to the circuit-level design and training of large-scale analog energy-based systems incorporating realistic nonlinear devices.