Closing the Design Loop: Bayesian Optimization Assisted Hierarchical Analog Layout Synthesis
Closing the Design Loop: Bayesian Optimization Assisted Hierarchical Analog Layout Synthesis
复制标题
DOI:
10.1109/dac18072.2020.9218621
复制
发表时间:
2020-07
期刊:
影响因子:
--
通讯作者:
Mingjie Liu;Keren Zhu;Xiyuan Tang;Biying Xu;Wei Shi;Nan Sun;D. Pan
中科院分区:
文献类型:
--
作者:
Mingjie Liu;Keren Zhu;Xiyuan Tang;Biying Xu;Wei Shi;Nan Sun;D. Pan
Existing analog layout synthesis tools provide little guarantee to post layout performance and have limited capabilities of handling system-level designs. In this paper, we present a closed-loop hierarchical analog layout synthesizer, capable of handling system designs. To ensure system performance, the building block layout implementations are optimized efficiently, utilizing post layout simulations with multi-objective Bayesian optimization. To the best of our knowledge, this is the first work demonstrating success in automated layout synthesis on generic analog system designs. Experimental results show our synthesized continuous-time ∆Σ modulator (CTDSM) achieves post layout performance of 65.9dB in signal to noise and distortion ratio (SNDR), compared with 67.8dB in the schematic design.