A delay-constrained optimization framework for low-power VLSI interconnect design using mathematical signal models

IF 2.5 4区工程技术 Q3 ENGINEERING, ELECTRICAL & ELECTRONIC

Journal of Computational Electronics Pub Date : 2025-09-22 DOI:10.1007/s10825-025-02416-0

V. Rajkumar, R. Amutha

引用次数: 0

Abstract

As VLSI technology scales to sub-7 nm nodes, interconnect-related delay and power dissipation become dominant design bottlenecks. This paper presents a comprehensive mathematical framework for modeling and optimizing interconnects in very-large-scale integration (VLSI) systems under delay constraints. Leveraging signal processing theory and circuit-level modeling, we introduce an enhanced delay model incorporating Elmore delay, crosstalk effects, and capacitive coupling. A constrained optimization strategy using Lagrangian relaxation and Karush–Kuhn–Tucker conditions is applied to minimize dynamic power while preserving signal integrity. Simulation results on 7 nm process technology demonstrate that the proposed method achieves up to 23% reduction in power with marginal delay overheads. Our framework is validated using HSPICE and Cadence Spectre on standard ISCAS85 and OpenCore benchmarks, providing a practical path to energy-efficient interconnect design.

Abstract Image

查看原文本刊更多论文

基于数学信号模型的低功耗VLSI互连设计延迟约束优化框架

随着VLSI技术扩展到7纳米以下节点，互连相关的延迟和功耗成为主要的设计瓶颈。本文提出了一个综合的数学框架，用于在延迟约束下对超大规模集成电路（VLSI）系统中的互连进行建模和优化。利用信号处理理论和电路级建模，我们介绍了一个包含Elmore延迟、串扰效应和电容耦合的增强延迟模型。采用拉格朗日松弛和Karush-Kuhn-Tucker条件的约束优化策略，在保持信号完整性的同时最小化动态功率。在7nm制程技术上的仿真结果表明，该方法在边际延迟开销的情况下，功耗降低了23%。我们的框架在标准ISCAS85和OpenCore基准测试上使用HSPICE和Cadence Spectre进行了验证，为节能互连设计提供了一条实用的途径。

本文章由计算机程序翻译，如有差异，请以英文原文为准。

求助全文

约1分钟内获得全文求助全文

来源期刊

Journal of Computational Electronics ENGINEERING, ELECTRICAL & ELECTRONIC-PHYSICS, APPLIED

CiteScore

4.50

自引率

4.80%

发文量

142

审稿时长

>12 weeks

期刊介绍： he Journal of Computational Electronics brings together research on all aspects of modeling and simulation of modern electronics. This includes optical, electronic, mechanical, and quantum mechanical aspects, as well as research on the underlying mathematical algorithms and computational details. The related areas of energy conversion/storage and of molecular and biological systems, in which the thrust is on the charge transport, electronic, mechanical, and optical properties, are also covered. In particular, we encourage manuscripts dealing with device simulation; with optical and optoelectronic systems and photonics; with energy storage (e.g. batteries, fuel cells) and harvesting (e.g. photovoltaic), with simulation of circuits, VLSI layout, logic and architecture (based on, for example, CMOS devices, quantum-cellular automata, QBITs, or single-electron transistors); with electromagnetic simulations (such as microwave electronics and components); or with molecular and biological systems. However, in all these cases, the submitted manuscripts should explicitly address the electronic properties of the relevant systems, materials, or devices and/or present novel contributions to the physical models, computational strategies, or numerical algorithms.