GPU-Accelerated Differentiable Framework Resolving Trade-off Between Speed and Accuracy in Switching Power Analysis (Duke, Synopsys)

Researchers from Duke University and Synopsys published a technical paper titled “DiffPower: GPU-Accelerated Differentiable Switching Power Analysis and Optimization.” Abstract Excerpt: The paper presents DiffPower as “a GPU-accelerated framework for differentiable power analysis and optimization. DiffPower translates design netlists into a PDK-agnostic bytecode representation, enabling analytical gradient computation via reverse-mode automatic differentiation, achieving up to a 1,002 × speedup