A new efficient waveform relaxation technique based on dynamically overlapped partitioning scheme is presented. This overlapped partitioning method enables the application of waveform relaxation technique to bipolar VLSI circuits. Instead of fixed overlapping, we select the depth of overlapping dynamically based on the sensitivity criteria. By minimizing the overlapped area, we could reduce the additional computational overhead which results from overlapping the partitions. This overlapped waveform relaxation method has better convergence properties due to smaller error introduced at each step compared with standard relaxation techniques. When overlapped partitioning is used in the case of digital circuits, the waveforms obtained after first iteration are nearly accurate. Therefore, by using these waveforms as initial guess waveforms for the second iterations we can reduce Newton-Raphson iterations at each time point.
The copyright of the original papers published on this site belongs to IEICE. Unauthorized use of the original or translated papers is prohibited. See IEICE Provisions on Copyright for details.
Copy
Vijaya Gopal BANDI, Hideki ASAI, "Dynamically Overlapped Partitioning Technique to Implement Waveform Relaxation Simulation of Bipolar Circuits" in IEICE TRANSACTIONS on Fundamentals,
vol. E77-A, no. 6, pp. 1080-1084, June 1994, doi: .
Abstract: A new efficient waveform relaxation technique based on dynamically overlapped partitioning scheme is presented. This overlapped partitioning method enables the application of waveform relaxation technique to bipolar VLSI circuits. Instead of fixed overlapping, we select the depth of overlapping dynamically based on the sensitivity criteria. By minimizing the overlapped area, we could reduce the additional computational overhead which results from overlapping the partitions. This overlapped waveform relaxation method has better convergence properties due to smaller error introduced at each step compared with standard relaxation techniques. When overlapped partitioning is used in the case of digital circuits, the waveforms obtained after first iteration are nearly accurate. Therefore, by using these waveforms as initial guess waveforms for the second iterations we can reduce Newton-Raphson iterations at each time point.
URL: https://global.ieice.org/en_transactions/fundamentals/10.1587/e77-a_6_1080/_p
Copy
@ARTICLE{e77-a_6_1080,
author={Vijaya Gopal BANDI, Hideki ASAI, },
journal={IEICE TRANSACTIONS on Fundamentals},
title={Dynamically Overlapped Partitioning Technique to Implement Waveform Relaxation Simulation of Bipolar Circuits},
year={1994},
volume={E77-A},
number={6},
pages={1080-1084},
abstract={A new efficient waveform relaxation technique based on dynamically overlapped partitioning scheme is presented. This overlapped partitioning method enables the application of waveform relaxation technique to bipolar VLSI circuits. Instead of fixed overlapping, we select the depth of overlapping dynamically based on the sensitivity criteria. By minimizing the overlapped area, we could reduce the additional computational overhead which results from overlapping the partitions. This overlapped waveform relaxation method has better convergence properties due to smaller error introduced at each step compared with standard relaxation techniques. When overlapped partitioning is used in the case of digital circuits, the waveforms obtained after first iteration are nearly accurate. Therefore, by using these waveforms as initial guess waveforms for the second iterations we can reduce Newton-Raphson iterations at each time point.},
keywords={},
doi={},
ISSN={},
month={June},}
Copy
TY - JOUR
TI - Dynamically Overlapped Partitioning Technique to Implement Waveform Relaxation Simulation of Bipolar Circuits
T2 - IEICE TRANSACTIONS on Fundamentals
SP - 1080
EP - 1084
AU - Vijaya Gopal BANDI
AU - Hideki ASAI
PY - 1994
DO -
JO - IEICE TRANSACTIONS on Fundamentals
SN -
VL - E77-A
IS - 6
JA - IEICE TRANSACTIONS on Fundamentals
Y1 - June 1994
AB - A new efficient waveform relaxation technique based on dynamically overlapped partitioning scheme is presented. This overlapped partitioning method enables the application of waveform relaxation technique to bipolar VLSI circuits. Instead of fixed overlapping, we select the depth of overlapping dynamically based on the sensitivity criteria. By minimizing the overlapped area, we could reduce the additional computational overhead which results from overlapping the partitions. This overlapped waveform relaxation method has better convergence properties due to smaller error introduced at each step compared with standard relaxation techniques. When overlapped partitioning is used in the case of digital circuits, the waveforms obtained after first iteration are nearly accurate. Therefore, by using these waveforms as initial guess waveforms for the second iterations we can reduce Newton-Raphson iterations at each time point.
ER -