PETA-Gmin:求解非线性电路的动态延拓算法

PETA-Gmin:dynamic continuation algorithm for solving nonlinear circuits

在集成电路设计的晶体管级电路仿真中,为了成功求解直流工作点,诞生了各种各样的直流延拓算法,如Gmin步进法、源步进法、伪瞬态法和同伦法等。随着集成电路的不断发展,对于现今具有强非线性的大规模电路,直流分析算法的收敛性能和效率仍然是一个巨大的挑战。针对具有强非线性的大规模电路在仿真过程中存在的收敛性较差以及仿真效率较低等问题,提出了一种名为PETA-Gmin的动态直流混合延拓算法。不同于之前所有的传统算法,PETA-Gmin将伪瞬态的过程成功地结合到了Gmin步进法中。由于伪电容的接入使得节点电压可以连续变化,从而有效地改善了仿真中解曲线不连续的问题,提高了收敛性。同时,Gmin步进法的快速步进保证了高效的仿真速率。目前,通过对工业界的大规模电路进行仿真,验证了所提PETA-Gmin算法的性能。实验结果表明,所提算法在仿真收敛性方面相对于Gmin步进法得到了显著提升,并且在仿真效率方面相对于最新的伪瞬态算法,其加速比最高(平均)可达4.89倍(2.57倍)。

In the transistor-level circuit simulation of integrated circuit design,in order to successfully solve the DC operating point,various DC continuation algorithms were born,such as Gmin stepping algorithm,source stepping algorithm,pseudo-transient algorithm,homotopy algorithm,etc.With the continuous development of integrated circuits,the convergence performance and efficiency of DC analysis algorithms are still a huge challenge for today’s large-scale circuits with strong nonlinearity.In order to solve the problems of poor convergence and low simulation efficiency in the simulation process of large-scale circuits with strong nonlinearities,a hybrid multi-phase continuation method named PsEudo-TrAnsientGmin stepping(PETA-Gmin)is proposed to improve the convergence while maintaining a high efficiency.Different from all conventional methods,the PETA-Gmin successfully integrates the pseudo-transient process into the Gmin stepping method.The pseudo-capacitor allows the node voltage to change continuously,thereby effectively improving the problem of discontinuity of the solution curve in the simulation,and improving the convergence.At the same time,the fast stepping of Gmin ensures an excellent simulation efficiency.The proposed PETA-Gmin approach is verified by industrial-level large-scale post-layout circuits.Experimental results demonstrate that the convergence is greatly improved compared with the Gmin stepping method,and that the approach achieves up to 4.89 X(2.57 X on average)speedup compared with the state-of-the-art pseudo transient analysis method.