A force directed partitioning algorithm for 3D floorplanning

Linquan Lyu, Takeshi Yoshimura

研究成果: Conference contribution

1 引用 (Scopus)

抜粋

Partitioning is very important to 3D IC physical design. Currently, hMetis (most famous partitioning software package) is widely used for partitioning. However, hMetis is for partitioning of hyper graphs. When the method is used for layer assignment of 3D floorplan/placement, some net might be too long. This paper develops a partitioning algorithm for 3D floorplan. The proposed combines a cost-based heuristic and force directed algorithm, which places the nodes considering attractive force and repulsive force, in order to solve the long net problem. We consider both the wire length and the number of TSVs in this work. The second part of the algorithm has two kinds according to the different emphasis. The experimental results demonstrated our algorithms can effectively reduce the wire length and the number of TSV. By comparing the experimental results of two focuses, we can get the differences and strengths of the two algorithms separately.

元の言語English
ホスト出版物のタイトルProceedings - 2017 IEEE 12th International Conference on ASIC, ASICON 2017
出版者IEEE Computer Society
ページ718-721
ページ数4
2017-October
ISBN(電子版)9781509066247
DOI
出版物ステータスPublished - 2018 1 8
イベント12th IEEE International Conference on Advanced Semiconductor Integrated Circuits, ASICON 2017 - Guiyang, China
継続期間: 2017 10 252017 10 28

Other

Other12th IEEE International Conference on Advanced Semiconductor Integrated Circuits, ASICON 2017
China
Guiyang
期間17/10/2517/10/28

ASJC Scopus subject areas

  • Hardware and Architecture
  • Electrical and Electronic Engineering

フィンガープリント A force directed partitioning algorithm for 3D floorplanning' の研究トピックを掘り下げます。これらはともに一意のフィンガープリントを構成します。

  • これを引用

    Lyu, L., & Yoshimura, T. (2018). A force directed partitioning algorithm for 3D floorplanning. : Proceedings - 2017 IEEE 12th International Conference on ASIC, ASICON 2017 (巻 2017-October, pp. 718-721). IEEE Computer Society. https://doi.org/10.1109/ASICON.2017.8252576