2020•Unpublished venueRequires access

Standard-Cell Scaling Framework with Guaranteed Pin-Accessibility

Chung‐Kuan Cheng, Daeyeal Lee, Dong-Won Park

Open publisher page 10 citations

Abstract

With the scaling of VLSI technologies, the design-technology co-optimization (DTCO) requires prompt development of standard cell libraries to explore scaling effects of various cell architectures. However, standard cell layout design demands holistic efforts for processing transistor placement and in-cell routing due to the limited routing tracks and complicated design rules. Thus, an automatic design framework of standard cell layout became essential in the advanced scaling. Conventional heuristic/divide-and-conquer approaches lack the optimality of solutions because of the limited solution space. In this paper, we propose a novel standard cell scaling framework that simultaneously finds an optimal solution in placement and routing with the pin-accessibility. To ensure the minimum number of pin-access points, we devise strict Boolean counter-based design constraints. We validate our framework using scaling parameters and cell architectures across sub-7nm technology nodes.

About this research paper

What this paper is about

With the scaling of VLSI technologies, the design-technology co-optimization (DTCO) requires prompt development of standard cell libraries to explore scaling effects of various cell architectures. However, standard cell layout design demands holistic efforts for processing transistor placement and in-cell routing due to the limited routing tracks and complicated design rules. Thus, an automatic design framework of standard cell layout became essential in the advanced scaling. Conventional heuristic/divide-and-conquer approaches lack the optimality of solutions because of the limited solution space. In this paper, we propose a novel standard cell scaling framework that simultaneously finds an optimal solution in placement and routing with the pin-accessibility. To ensure the minimum number of pin-access points, we devise strict Boolean counter-based design constraints. We validate our framework using scaling parameters and cell architectures across sub-7nm technology nodes.

Why it matters

OpenAlex reports 10 citations for this work. Citation counts describe recorded attention and do not establish research quality.

Key contribution

A contribution statement is not available in the OpenAlex record.

Method / approach

Method details are not available in the OpenAlex metadata.

Main findings

Findings are not separately available in the OpenAlex metadata.

Limitations

Limitations are not available in the OpenAlex metadata.

Applications

Application details are not available in the OpenAlex metadata.

Available abstract

With the scaling of VLSI technologies, the design-technology co-optimization (DTCO) requires prompt development of standard cell libraries to explore scaling effects of various cell architectures. However, standard cell layout design demands holistic efforts for processing transistor placement and in-cell routing due to the limited routing tracks and complicated design rules. Thus, an automatic design framework of standard cell layout became essential in the advanced scaling. Conventional heuristic/divide-and-conquer approaches lack the optimality of solutions because of the limited solution space. In this paper, we propose a novel standard cell scaling framework that simultaneously finds an optimal solution in placement and routing with the pin-accessibility. To ensure the minimum number of pin-access points, we devise strict Boolean counter-based design constraints. We validate our framework using scaling parameters and cell architectures across sub-7nm technology nodes.

Key concepts: Standard cell, Routing (electronic design automation), Computer science, Very-large-scale integration, Scaling, Heuristic, Physical design, Electronic design automation

Related papers

Back to paper searchBrowse research topicsOriginal source
Standard-Cell Scaling Framework with Guaranteed Pin-Accessibility — Research Paper | ScholarLens