New preprint on variation-aware clock-mesh analysis with GNNs

Our new preprint, "A Surrogate Model for Variation-Aware Clock-Mesh Analysis and Design-Space Exploration using Graph Neural Networks" by Muhammad Hadir Khan, Wajid Ali, Arya Thottiparambil Manikandan, and Matthew R. Guthaus, is now on arXiv. The work introduces a graph neural network surrogate model of the clock mesh, trained on SPICE data, that predicts delay, slew, skew under on-chip variation, and power. It is far faster than the approximate methods used in current clock-mesh synthesis while being more accurate, enabling faster clock-mesh design-space exploration. Congratulations to the authors!