Technology tracking of non manhattan VLSI layout
DAC '89 Proceedings of the 26th ACM/IEEE Design Automation Conference
Over-the-cell channel routing for high performance circuits
DAC '92 Proceedings of the 29th ACM/IEEE Design Automation Conference
Rectilinear paths among rectilinear obstacles
Discrete Applied Mathematics
Manhattan or non-Manhattan?: a study of alternative VLSI routing architectures
GLSVLSI '00 Proceedings of the 10th Great Lakes symposium on VLSI
VISI Physical Design Automation: Theory and Practice
VISI Physical Design Automation: Theory and Practice
TWIGY - a topological algorithm based routing system
DAC '81 Proceedings of the 18th Design Automation Conference
DAC '78 Proceedings of the 15th Design Automation Conference
Multilevel full-chip routing for the X-based architecture
Proceedings of the 42nd annual Design Automation Conference
A Performance-Driven Multilevel Framework for the X-Based Full-Chip Router
Integrated Circuit and System Design. Power and Timing Modeling, Optimization and Simulation
PIXAR: A performance-driven X-architecture router based on a novel multilevel framework
Integration, the VLSI Journal
A parallel dual-scanline algorithm for partitioning parameterized 45-degree polygons
ACM Transactions on Design Automation of Electronic Systems (TODAES) - Special Section on Networks on Chip: Architecture, Tools, and Methodologies
| Hi-index | 0.00 |
The availability of multiple metal layers in modern IC processes raises the possibility of using non-Manhattan routing on some of the layers in order to reduce the average interconnect length, and thus improve performance and routability. In this paper, we present novel algorithms for both Manhattan and non-Manhattan multi-layer maze routing. The algorithms in principle can be extended to an arbitrary number of layers, but the paper focuses on four-layer routing, two in horizontal and two in vertical directions for Manhattan, and one layer each in horizontal, vertical, 45-degree and 135-degree directions for non-Manhattan routing. The non-Manhattan algorithms show an improvement of up to 12.2% in average wire length compared to Manhattan routing for two general MCNC benchmarks.