Clockless%20Computing - PowerPoint PPT Presentation

About This Presentation
Title:

Clockless%20Computing

Description:

dual-rail may be faster (if completion times vary widely) 2-phase vs. 4-phase ... rail needs fewer gates/wires/pins. design and verification effort. dual-rail, ... – PowerPoint PPT presentation

Number of Views:48
Avg rating:3.0/5.0
Slides: 15
Provided by: Montek5
Learn more at: http://www.cs.unc.edu
Category:

less

Transcript and Presenter's Notes

Title: Clockless%20Computing


1
Clockless Computing
  • Montek Singh
  • Thu, Aug 23, 2007

2
Preliminaries
  • How is data represented in an asynchronous
    system?
  • How is information exchanged? control
    signaling (handshake styles)

3
Data Encoding Bundled Data
  • Single-rail Bundled Datapath simplest approach
  • widely used
  • Features
  • datapath 1 wire per bit (e.g. standard sync
    blocks)
  • matched delay produces delayed done signal
  • worst-case delay longer than slowest path
  • Practical style can reuse sync components small
    area
  • Fixed (worst-case) completion time

4
Bundled Data Completion Sensing
  • Delay Matching
  • either single worst-case delay
  • or, fine-grain delay
  • Speculative completion
  • choose delay on the fly
  • start with shortest delay increase as needed

5
Data Encoding Dual-Rail
  • Dual-rail uses 2 wires per data bit

Each Dual-Rail Pair provides both data value and
validity
  • provides robust data-dependent completion
  • needs completion detectors

6
Dual-Rail Completion Sensing
  • Dual-Rail Completion Detector
  • combines dual-rail signals
  • indicates when all bits are valid (or reset)
  • C-element
  • if all inputs1, output ? 1
  • if all inputs0, output ? 0
  • else, maintain output value
  • OR together 2 rails per bit
  • Merge results using a Müller C-element

7
Handshaking Styles 4-phase
  • 4-Phase requires 4 events per handshake
  • Level-sensitive ? simpler logic implementation
  • Overhead of return-to-zero (RTZ or resetting)
  • extra events which do no useful computation

8
Handshaking Styles 2-phase
  • 2-Phase requires 2 events per handshake
  • a.k.a. transition signaling
  • Elegant no return-to-zero
  • Slower logic implementation
  • logic primitives are inherently level-sensitive,
    not event-based (at least in CMOS)

9
Handshaking Styles Pulse Mode
  • Pulse Mode combines benefits of 2-phase and
    4-phase
  • use pulses to represent events

start next event
start event
Request
next event done
event done
Acknowledge
  • No return-to-zero (like 2-phase)
  • Level-based implementation (like 4-phase)
  • Need a timing constraint on pulse width

10
Handshaking Styles Single-Track
  • Single-Track combines req and ack onto single
    wire!
  • one wire used for bidirectional communication
  • sender raises, receiver lowers
  • Efficient protocol no return-to-zero,
    level-based
  • Need aggressive low-level design techniques
  • much effort to ensure reliability, satisfy timing
    constraints

11
Handshaking Data Representation
  • Several combinations possible
  • dual-rail 4-phase, single-rail 4-phase, dual-rail
    2-phase, and single-rail 2-phase
  • Example dual-rail 4-phase

A
B
  • dual-rail data functions as an implicit
    request
  • 4-phase cycle between acknowledge and implicit
    request

12
Other Data Representation Styles
  • Level-Encoded Dual-Rail (LEDR)
  • 2 wires per bit data and phase
  • exactly one wire per bit changes value
  • if new value is different, data wire changes
    value
  • else phase wire change value
  • M-of-N Codes
  • N wires used for a data word
  • M wires (M lt N) change value
  • Values of N and M have impact on
  • information transmitted, power consumed and logic
    complexity
  • Knuth codes, Huffman codes,

13
Which to use?
  • Depends on several performance parameters
  • speed
  • single-rail vs. dual-rail
  • single-rail may be faster (if designed
    aggressively)
  • dual-rail may be faster (if completion times vary
    widely)
  • 2-phase vs. 4-phase
  • 2-phase may be faster (if logic overhead is
    small)
  • 4-phase may be faster (if overhead of
    return-to-zero is small)
  • power consumption
  • 2-phase typically has fewer gate transitions (?
    lower power)
  • amount of logic used (gates/wires/pins ? chip
    area)
  • single-rail needs fewer gates/wires/pins
  • design and verification effort
  • dual-rail, 1-of-N, M-of-N, Knuth codes
  • delay-insensitive robust in the presence of
    arbitrary delays
  • single-rail requires greater timing
    verification effort

14
Homework 2 (due Thu Aug 30)
  • Suppose you are given N wires
  • Which M-of-N encoding (i.e. what M) encodes most
    information?
  • Suppose you have to encode 4-bit values
  • Which M-of-N encoding yields fewest wires?
  • Suppose you can switch at most 2 wires
  • Which M-of-N encoding yields fewest wires for
    4-bit values?
Write a Comment
User Comments (0)
About PowerShow.com