Amortized Neural Optimization for Pre-Layout Signal Integrity Design Space Exploration using Differentiable Surrogates

2026-06-05

AI Summary

This work addresses the inefficiency of design space exploration for high-speed signal integrity during pre-layout stages, which suffers from the prohibitive computational cost of multi-corner simulations and iterative optimization. To overcome this challenge, the authors propose Amortized Neural Optimization (ANO), introducing amortized optimization to this domain for the first time. ANO employs a fully differentiable neural surrogate model trained offline to learn an end-to-end optimization policy, enabling global optimization via analytical gradients and amortizing inference into a single forward pass—thereby eliminating online iterations entirely. Evaluated on tasks including DDR5 DFE tuning, 9-dimensional SerDes equalization, and DDR3 differential pair routing, ANO achieves speedups of three to four orders of magnitude over conventional methods, reducing multi-corner optimization across hundreds of thousands of instances from days to milliseconds at the cost of only ~10% optimality loss.

Abstract

Pre-layout design space exploration (DSE) for high-speed signal integrity (SI) analysis is often limited by the computational cost of simulations and iterative optimization algorithms within modern electronic design automation (EDA) workflows. While machine learning surrogate models accelerate the simulation step, optimizing designs still requires utilizing iterative black-box search methods. This iterative nature scales poorly, making multi-corner sweeps computationally expensive. As a solution, this paper proposes amortized neural optimization (ANO) for pre-layout SI design. ANO entirely eliminates iterative black-box inference by utilizing fully differentiable neural network surrogate models. ANO extracts analytical gradients from the surrogate to train a global optimization policy. Instead of solving the optimization problem repeatedly at inference, the optimization process is learned offline and therefore amortized. Once the ANO policy is trained, it maps different channel contexts directly to near-optimal design parameters in a single deterministic forward pass. The efficiency and accuracy of the ANO framework are demonstrated based on three complex SI design scenarios, including DDR5 decision feedback equalization (DFE), 9-dimensional SerDes Tx/Rx co-equalization, and DDR3 DQS differential pair routing to optimize eye diagram metrics under intra-pair skew constraints. By trading roughly 10% in optimality compared to instance-specific black-box algorithms, it realizes speedups of three to four orders of magnitude. For a large-scale 320,000-instance multi-corner SerDes sweep optimization, ANO collapses what would have taken days of computation using iterative search algorithms into a single batched forward pass that completes in milliseconds. This transforms computationally expensive SI optimization into real-time and interactive pre-layout DSE.

Problem

• signal integrity
• design space exploration
• pre-layout optimization
• computational cost
• iterative optimization

Innovation

• amortized neural optimization
• differentiable surrogate
• pre-layout design space exploration
• signal integrity
• real-time optimization