High-speed QDI asynchronous pipelines

Abstract

This paper introduces two new high-speed quasi delay insensitive (QDI) asynchronous pipeline templates. These new high throughput templates support complex non-linear pipeline structures and are well suited for fine-grain pipelining. Timing analysis and HSPICE simulations show that these templates are 20% and 40% faster than known QDI counterparts.

Assessment of nano-scale Muller C-elements under variability based on a new fault model

“…To have a more fair discussion, we investigate the area tradeoff of using different C-element implementations in a well-known asynchronous template called Pre-Charged Full Buffer (PCFB) template [17]Error! Reference source not found.…”

MOUSETRAP: High-Speed Transition-Signaling Asynchronous Pipelines

“…While several distinct pipeline styles are introduced, the most commonly-used ones have cycle times that are 18 gate delays or greater. Other QDI styles include those by Ozdag and Beerel [24], which improve upon the performance of [18] and [20] through use of concurrency reduction to simplify the design of the dynamic function blocks (i.e., reduced stack size). Finally, Ferretti and Beerel [9] combine the robustness of delay-insensitive encoding with the effiency of the single-track handshaking of [33] and introduce single-track 1-of-pipelines.…”

Heterogeneous Latch-Based Asynchronous Pipelines

“…Fine-grained pipeline structures were introduced by Williams [30], and improved on by many others [18,15,17,13,32,16,22]. These pipeline templates are usually implemented using dynamic logic, which provides intrinsic storage capability, but are not amenable to automatic synthesis with commercial standard-cell CAD tools.…”