Adapting Asynchronous Circuits to Operating Conditions by Logic - - PowerPoint PPT Presentation

adapting asynchronous circuits
SMART_READER_LITE
LIVE PREVIEW

Adapting Asynchronous Circuits to Operating Conditions by Logic - - PowerPoint PPT Presentation

Adapting Asynchronous Circuits to Operating Conditions by Logic Parameterisation Andrey Mokhov, Danil Sokolov, Alex Yakovlev Newcastle University, UK Designers dilemma Contradicting requirements: o performance, energy consumption,


slide-1
SLIDE 1

Adapting Asynchronous Circuits to Operating Conditions by Logic Parameterisation

Andrey Mokhov, Danil Sokolov, Alex Yakovlev Newcastle University, UK

slide-2
SLIDE 2
  • Contradicting requirements:
  • ↑ performance, ↓ energy consumption, ↑ robustness
  • high performance → high energy consumption
  • low energy consumption → low robustness
  • high robustness → low performance
  • Competing implementation styles:

– Synchronous – Asynchronous

  • Delay Insensitive (DI)
  • Speed Independent (SI) or Quasi Delay Insensitive (QDI)
  • Bundled data, relative timing assumptions (TA)
  • ...

Designer’s dilemma

slide-3
SLIDE 3

Energy-efficiency v Robustness

slide-4
SLIDE 4

Parameterised Circuits

  • Static parameters set at test/binning stages
  • Dynamic parameters:

– Power management controller – Maintenance mechanisms

slide-5
SLIDE 5

Parameterised Circuits: Trivial Approach

  • Easy to design!
  • Large overheads
  • How can we do better?
slide-6
SLIDE 6
  • Goal:

– Combine circuit implementations efficiently

  • Key observations:

– Externally: the circuits have the same interface – Internally: the circuits behave similarly

  • Solution: use a model that can capture

functional similarities at the circuit level

– Conditional Partial Order Graphs almost fit

Parameterised Circuits: Better Approach

slide-7
SLIDE 7

Conditional Partial Order Graphs

x=0 y=0 x=1 y=0 x=0 y=1 x=1 y=1

slide-8
SLIDE 8

Conditional Partial Order Graphs

 Describe concurrency and causality  Capture similarities in behaviours  Represent families of partial orders

  • acyclic
  • not directly applicable to circuit specification

 Is it possible to specify cyclic behaviour using acyclic objects?

–Yes!

slide-9
SLIDE 9

Describing cycles

slide-10
SLIDE 10

Describing cycles

slide-11
SLIDE 11

Describing cycles

slide-12
SLIDE 12

Describing cycles

slide-13
SLIDE 13

Describing cycles

x = x = 1

slide-14
SLIDE 14

Describing cycles

Initial state: x = 1 set x = 0 set x = 1 x – x +

slide-15
SLIDE 15

Describing cycles

x = 1

 

slide-16
SLIDE 16

Describing cycles

x = 0

   

slide-17
SLIDE 17

Describing cycles

x = 1

    

slide-18
SLIDE 18

Describing circuits: oscillator

x = 0 x = 0

STG Conditional Signal Graphs (CSGs)

slide-19
SLIDE 19

Describing circuits: C-element

STG

slide-20
SLIDE 20

Describing circuits: C-element

CSG

slide-21
SLIDE 21

Describing circuits: C-element

slide-22
SLIDE 22

Describing circuits: C-element

slide-23
SLIDE 23

Describing circuits: C-element

slide-24
SLIDE 24

Describing circuits: C-element

slide-25
SLIDE 25

Describing circuits: C-element

slide-26
SLIDE 26

Describing circuits: C-element

slide-27
SLIDE 27

Describing circuits: C-element

slide-28
SLIDE 28

Describing circuits: C-element

slide-29
SLIDE 29

Circuit composition

C-element AND-gate C/AND-element

slide-30
SLIDE 30

Circuit composition

C-element: ↑ max(a↑, b↑) ↓ max(a↓, b↓) AND-gate: ↑ max(a↑, b↑) ↓ min(a↓, b↓) C/AND-element: can be chosen in run-time!

slide-31
SLIDE 31

Circuit composition

  • Algebraic operations with CSGs:

– Addition (overlay): G1 + G2 – Scalar multiplication (encoding): f  G

  • Composition of C-element and AND-gate:
  • General composition of n circuits:
  • Fast structural operation (if encoding is given)
slide-32
SLIDE 32

Design flow

  • Heterogeneous models

and implementation styles

  • Not computationally

expensive:

– composition is fast – exact encoding is slow but there are fast approximate methods

slide-33
SLIDE 33

Example: Read/Write controller

slide-34
SLIDE 34

Example: possible implementations

Delay Insensitive (DI) T = d + max{d1, d2, d3} + C3 Partial Acknowledgement (PA) T = d + max{d1, d2} + C2 Timing Assumptions (TA) T = d + I Synchronous (CL) T = tclock

slide-35
SLIDE 35

Example: parameterised controller

Can be switched off by power-gating in TA/CL modes

slide-36
SLIDE 36

Example: simulation results

slide-37
SLIDE 37
  • Trivial approach:

– Latency overhead: 120-182%, 122% on average – Energy overhead: 151-330%, 201% on average

  • Our approach:

– Latency overhead: 0-40%, 19% on average – Energy overhead: 0-98%, 23% on average

  • Latency/energy savings at the system level!

Cost of Parameterisation

slide-38
SLIDE 38
  • “Don’t oppose different implementation styles,

take advantage of their benefits!” (reviewer X)

  • New model for circuit-level description and

composition

  • Cost of parameterisation is not prohibitive
  • Future work:

– Tool prototyping – CSC signals awareness – Power-gating mechanisms

Conclusions & Future work

slide-39
SLIDE 39

Questions?