Abstract
Analog IC designers and design houses have been accumulating their own design knowledge and constructing their own analog design repositories, including various design specifications, applications, and process technologies. As most of the analog layouts are handcrafted art works, different designers/companies may have different layout guidelines and preferences. When generating a new layout for certain analog design which already exists or is similar to any of those in the repositories, but with different circuit parameters or process technology files, applying layout migration is usually more preferable than starting from scratch. This paper introduces a holistic framework and new layout generation methodology to achieve analog layout integrity through learning and migration. The introduced methodology can effectively and efficiently preserve the preferences of placement and routing topologies from legacy layouts to new ones.
Original language | English |
---|---|
Article number | 9256451 |
Journal | IEEE/ACM International Conference on Computer-Aided Design, Digest of Technical Papers, ICCAD |
Volume | 2020-November |
DOIs | |
State | Published - 2 Nov 2020 |
Event | 39th IEEE/ACM International Conference on Computer-Aided Design, ICCAD 2020 - Virtual, San Diego, United States Duration: 2 Nov 2020 → 5 Nov 2020 |
Keywords
- Analog layout
- analog placement
- analog routing
- building block
- circuit classification
- circuit identification
- circuit structure analysis
- layout pattern
- layout template
- machine learning
- migration