Can I read Efficient Static Timing Analysis and Applications Using Edge Masks on EtoBox?

Efficient Static Timing Analysis and Applications Using Edge Masks by Mike Hutton; David Karchmer; Bryan Archell; Jason Govig is a scholarly article available to read on EtoBox.

What is Efficient Static Timing Analysis and Applications Using Edge Masks about?

Static timing analysis (STA) with multiple clock domains and complicated exception conditions is a complex practical problem that can dramatically increase compilation time, both for back-end analysis and during place and route. In FPGA placement, timing analysis with many constraints can dominate placement run-time. In this paper we introduce a simple binary edge-mask data structure on arcs in a timing netlist which allows for efficient timing analysis in the presence of many such constraints. The technique applies to either BFS or DFS-based timing analysis. Preliminary implementations on just the basic concept show a 59% decrease in STA run-time for multi-clock designs, indicating that significant benefit is to be gained from a complete implementation. On a set of heavily constrained designs this benefit improved to 80% run-time decrease. Further applications of the edge-mask concept are shown to efficiently deal with thru-x constraints, enumerating the k-longest paths in a timing graph, and partial/incremental timing analysis aimed at significant improvements in placement time.

Author

Mike Hutton; David Karchmer; Bryan Archell; Jason Govig

Publisher

ACM

Published

2005

Language

EN