High-Performance Asynchronous Pipelines: An Overview

@article{Nowick2011HighPerformanceAP,
  title={High-Performance Asynchronous Pipelines: An Overview},
  author={Steven M. Nowick and Montek Singh},
  journal={IEEE Design \& Test of Computers},
  year={2011},
  volume={28},
  pages={8-22}
}
Pipelining is a key element of high-performance design. Distributed synchronization is at the same time one of the key strengths and one of the major difficulties of asynchronous pipelining. It automatically provides elasticity and on-demand power consumption. This tutorial provides an overview of the best-in-class asynchronous pipelining methods that can be used to fully exploit the advantages of this design style, covering both static and dynamic logic implementations. 

Figures from this paper

Performance-Driven Clustering of Asynchronous Circuits
TLDR
This method provides a form of automatic pipelined in which the throughput of the overall design is not limited to the clock frequency or the level of pipelining in the original Register-Transfer Level (RTL) specification.
Asynchronous Design—Part 1: Overview and Recent Advances
TLDR
The first part of the two-part article on asynchronous design presents a chronicle of past and recent commercial advances, as well as technical foundations, and highlights the enabling role of asynchronous design in two application areas: GALS systems and networks-on-chip.
Implementation of asynchronous pipeline using Transmission Gate logic
TLDR
Sutherland Pipeline is implemented, a static asynchronous pipeline using transmission gate logic family, which aims to implement these pipelines more efficiently by reducing their area overhead and dynamic power dissipation.
Reconfigurable asynchronous pipelines: From formal models to silicon
TLDR
This paper presents a model and an open-source tool for the design and verification of reconfigurable asynchronous pipelines, and validates this approach in silicon.
A New Synchronous circuit for Elastic Pipeline Architecture
TLDR
A simple pipeline circuit for Synchronous design is presented that implements the implementation of elastic communication channels and is easily implementable in both ASIC and FPGA independent of the EDA (Electronic Design Automation).
High speed dynamic asynchronous pipeline: Self-Controlled approach
This paper proposes a high speed asynchronous pipeline architecture, called Self-Controlled (SC) pipeline, targeting dynamic data path. The SC pipeline uses Williams' PSO pipeline as the base
A Survey on Efficient Low Power Asynchronous Pipeline Design Based on the Data Path Logic
TLDR
A survey on high-throughput and ultra low-power asynchronous pipeline design method targeting to latch- free and extremely fine-grain design, which offers the pipeline low power consumption as well as high throughput by reducing the overhead problems.
Design and Implementation of Reconfigurable Asynchronous Pipelines
TLDR
This article presents a model and tool support for the design and verification of reconfigurable asynchronous pipelines, and designs and fabricates a test chip that demonstrates the benefits and costs of dynamic reconfigurability, as well as highlights the resilience of asynchronous pipelines.
High-Speed Dynamic Asynchronous Pipeline: Self-Precharging Style
TLDR
This brief proposes a new type of high throughput asynchronous pipeline structure called self-precharging (SP) pipeline that targets dynamic linear datapaths for fine grain and gate level pipelining and enables SP pipeline to deliver multi gigahertz throughput without degrading the per-stage forward latency of the pipeline.
...
...

References

SHOWING 1-10 OF 46 REFERENCES
Pipelined Asynchronous Circuits
TLDR
A design style for implementing communicating sequential processes (CSP) as quasi delay insensitive asynchronous circuits, based on the compilation method of [1], which can easily implement circuits with some slack between inputs and outputs is presented.
The Design of High-Performance Dynamic Asynchronous Pipelines: High-Capacity Style
  • Montek Singh, S. Nowick
  • Computer Science
    IEEE Transactions on Very Large Scale Integration (VLSI) Systems
  • 2007
This paper introduces a high-throughput asynchronous pipeline style, called high-capacity (HC) pipelines, targeted to datapaths that use dynamic logic. This approach includes a novel
MOUSETRAP: High-Speed Transition-Signaling Asynchronous Pipelines
  • Montek Singh, S. Nowick
  • Computer Science
    IEEE Transactions on Very Large Scale Integration (VLSI) Systems
  • 2007
TLDR
An asynchronous pipeline style is introduced for high-speed applications, called MOUSETRAP, which uses standard transparent latches and static logic in its datapath, and small latch controllers consisting of only a single gate per pipeline stage to handle more complex system architectures.
The Design of High-Performance Dynamic Asynchronous Pipelines: Lookahead Style
  • Montek Singh, S. Nowick
  • Computer Science
    IEEE Transactions on Very Large Scale Integration (VLSI) Systems
  • 2007
TLDR
A new class of asynchronous pipelines is proposed, called lookahead pipelines (LP), which use dynamic logic and are capable of delivering multi-gigahertz throughputs and are characterized by low-cost control structures and the avoidance of explicit latches.
Speculative completion for the design of high-performance asynchronous dynamic adders
TLDR
This paper presents an in-depth case study in high-performance asynchronous adder design that uses single-rail bundled datapaths but also allows early completion, and introduces five new dynamic designs for Brent-Kung and Carry-Bypass adders.
Multisynchronous and Fully Asynchronous NoCs for GALS Architectures
TLDR
Two approaches are described to implementing a distributed NoC in a GALS environment to address the difficulty of distributing a synchronous clock signal on the entire chip area.
Robust interfaces for mixed-timing systems
  • T. Chelcea, S. Nowick
  • Computer Science
    IEEE Transactions on Very Large Scale Integration (VLSI) Systems
  • 2004
TLDR
This paper presents several low-latency mixed-timing FIFO (first-in-first-out) interfaces designs that interface systems on a chip working at different speeds and initial simulations for both latency and throughput are promising.
Performance Analysis of Asynchronous Circuits Using Markov Chains
TLDR
Asynchronous systems generally consist of a collection of parallel computation processes that synchronize and exchange data through handshaking-based communication that facilitates pipelining communication across long distances, optimizing for average case behavior, and integrating slow and fast processes.
Low Power and Energy Efficient Asynchronous Design
TLDR
This paper surveys the most promising low-power and energy-efficient asynchronous design techniques that can lead to substantial advantages over synchronous counterparts, including microprocessors, application specific designs, and networks on chip.
An asynchronous NOC architecture providing low latency service and its multi-level design framework
TLDR
The proposed NOC protocol and its asynchronous implementation are presented as well as the multi-level modeling approach using SystemC language and transaction-level-modeling, which shows that the asynchronous NOC can offer 5 Gbytes/s throughput in a 0.13 /spl mu/m CMOS technology.
...
...