A Mini-Review of Methods for Forecasting Post Routing Congestion from Placement in FPGA Physical Design

Umair F. Siddiqi, Sadiq M. Sait

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

Abstract

Placement and routing are two time-consuming steps in the FPGA physical design flow and can take hours or even days. The placement steps map the logic elements of the netlist onto the computational resources of the FPGA, and the routing step is responsible for finding paths to connect the logic blocks through the FPGA routing fabric. The routing resources have a fixed capacity; using them beyond their capacity results in congestion. Congestion can cause routing failures and make it difficult to achieve timing closure and the placement algorithms should produce solutions that do not generate congestion in routing. Researchers have employed many approaches, from elementary ways (e.g., counting the number of pins) to accurate artificial intelligence-driven methods. This article categorizes congestion estimation methods developed over the last 25 years, providing an overview and survey of methods belonging to each category. To aid in understanding, simple experiments are included to illustrate each method's application. Finally, the article concludes by outlining potential future research directions in the field.

Original languageEnglish
Title of host publication2025 6th International Conference on Artificial Intelligence, Robotics, and Control, AIRC 2025
PublisherInstitute of Electrical and Electronics Engineers Inc.
Pages281-290
Number of pages10
ISBN (Electronic)9798331543488
DOIs
StatePublished - 2025
Event6th International Conference on Artificial Intelligence, Robotics, and Control, AIRC 2025 - Savannah, United States
Duration: 7 May 20259 May 2025

Publication series

Name2025 6th International Conference on Artificial Intelligence, Robotics, and Control, AIRC 2025

Conference

Conference6th International Conference on Artificial Intelligence, Robotics, and Control, AIRC 2025
Country/TerritoryUnited States
CitySavannah
Period7/05/259/05/25

Bibliographical note

Publisher Copyright:
© 2025 IEEE.

Keywords

  • FPGA
  • congestion prediction
  • generative AI
  • machine and deep learning
  • placement
  • routing

ASJC Scopus subject areas

  • Mechanical Engineering
  • Control and Optimization
  • Artificial Intelligence
  • Computer Science Applications

Fingerprint

Dive into the research topics of 'A Mini-Review of Methods for Forecasting Post Routing Congestion from Placement in FPGA Physical Design'. Together they form a unique fingerprint.

Cite this