Don't-Care Aware ESOP Extraction via Reduced Decomposition-Tree Exploration
Chun-Yu Wei, Jie-Hong R. Jiang
摘要
Exclusive-OR Sum-of-Products expressions (ESOPs) are vital for circuit synthesis of arithmetic functions and emerging technologies. The state-of-the-art ESOP extraction methods are limited in their inefficient exhaustive exploration strategy, significant optimality loss in the divide-and-conquer process, and incapability of handling don’t-cares. This work overcomes these limitations by reduced decomposition exploration with a cost estimation and refinement strategy, generally applicable to incompletely specified functions. Experiments show up to 29× (average 12×) runtime and 28× (average 13×) memory-usage improvements with a quality close to the exact optimum obtained by exhaustive exploration for completely-specified functions, and substantial ESOP simplification with don’t-cares for incompletely-specified functions.
问问这篇 Paper
问问你的智能体。
Lune 读过与它相关的顶会 Paper,每个回答都会注明依据哪几篇。
相关 Paper
- Search Strategies for Topological Network OptimizationMichael D. MoffittAAAI 2022
- Search space characterization for approximate logic synthesisLinus Witschen, Tobias Wiersema, Lucas Reuter, Marco PlatznerDAC 2022 · 被引用 3 次
- An Efficient Asynchronous Batch Bayesian Optimization Approach for Analog Circuit SynthesisShuhan Zhang, Fan Yang, Dian Zhou, Xuan ZengDAC 2020 · 被引用 41 次
- Vectorization for digital signal processors via equality saturationAlexa VanHattum, Rachit Nigam, Vincent T. Lee, James Bornholt 等ASPLOS 2021 · 被引用 57 次
- Circuit Minimization with QBF-Based Exact SynthesisFranz-Xaver Reichl, Friedrich Slivovsky, Stefan SzeiderAAAI 2023 · 被引用 14 次
