Timing-Driven Placement

  • Majid Sarrafzadeh
  • Maogang Wang
  • Xiaojian Yang
Chapter

Abstract

Recent work on wirelength optimization suggests that hierarchical [23] or multi-level [27, 161] approach is indispensable to efficiently solve large-scale placement problem. Correspondingly, timing-driven placement problem using hierarchical or multi-level framework should draw research attention. Several recent works address this problem. Ou et al. [114] adopt a net-cut control method in min-cut placement. Halpin et al. [70] presented a linear programming based net length control for recursive bisection placement. Kahng et al. [83] extended the placement flow in [23] to incorporate direct minimization of the critical path. The above approaches shows the trend of the combination between traditional timing-driven approach and top-down placement flow. However, delay budgeting, as a previously effective approach, has not been applied into the top-down placement framework.

Keywords

Delay Bound Placement Flow Delay Budget Design Hierarchy Delay Violation 
These keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.

Preview

Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.

Notes

  1. 1.
    Here we assume that the internal delay from different input pins to the output pin are the same. We will interchangeably use slack of node and slack of net in this paper.Google Scholar
  2. 2.
    It is also associated to the net which node vz drives, under the assumption that each node has one output and the net delay is calculated using the bounding box of the net.Google Scholar
  3. 3.
    The delay of a net is proportional to the half-perimeter of the bounding box of the net.Google Scholar
  4. 4.
    The node which drives that net will be assigned the weight.Google Scholar

Copyright information

© Springer Science+Business Media New York 2003

Authors and Affiliations

  • Majid Sarrafzadeh
    • 1
  • Maogang Wang
    • 2
  • Xiaojian Yang
    • 1
  1. 1.Computer Science DepartmentUniversity of CaliforniaLos AngelesUSA
  2. 2.Cadence Design Systems, Inc.USA

Personalised recommendations