Predicated Asynchronous Netlists for Unused Computation Control

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Asynchronous circuit designs consume power even when they produce unused computation results, leading to increased power consumption due to the inherent nature of their operation.

Innovation Solution

The conversion of non-predicated asynchronous netlists to predicated asynchronous netlists, where additional logic circuits dynamically turn off portions of the design that produce unused results, using tools and modules to simulate and synthesize the circuits and incorporate predication logic to control partially utilized portions.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Productivity

If asynchronous operations are used to improve processing speed and flexibility, then productivity is improved, but power consumption increases due to continued operation of unused computation portions

Engineering Contradiction:
Improveprocessing speedVSAvoidpower consumption
Core Design Contradiction:
ProductivityVSUse of energy by moving object

Solution Approach 1:

The patent applies dynamics by making the computation portions dynamically controllable through predication logic. The circuit transitions from a static always-on architecture to a dynamic one where computation portions are selectively enabled or disabled based on runtime conditions. This is achieved by introducing predicate signals that control the activation of computation portions, allowing the system to adapt its power consumption to actual computational needs while maintaining asynchronous operation benefits.

Inventive Principle:
Principle #15Dynamics

Solution Approach 2:

The patent applies local quality by enabling different portions of the circuit to have different operational states simultaneously. Instead of uniformly enabling or disabling the entire circuit, the predication logic allows specific computation portions to be activated or deactivated independently based on local conditions. This granular control ensures that only necessary computations consume power, reducing overall power consumption while maintaining productivity in active portions.

Inventive Principle:
Principle #3Local quality

2Use of energy by moving object

If predication logic is added to control computation portions, then power consumption is reduced by eliminating unused computations, but device complexity increases

Engineering Contradiction:
Improvepower consumptionVSAvoidcircuit complexity
Core Design Contradiction:
Use of energy by moving objectVSDevice complexity

Solution Approach 1:

The patent applies preliminary action by pre-computing predicate values that determine whether computation portions should be activated. The predication logic evaluates conditions in advance and generates control signals that prepare the circuit for selective activation. This preliminary evaluation allows the main computation portions to be controlled efficiently without adding complex real-time decision-making logic, thereby reducing power consumption with minimal increase in overall circuit complexity.

Inventive Principle:
Principle #10Preliminary action

Solution Approach 2:

The patent introduces predication logic as an intermediary layer between the control unit and computation portions. This intermediary evaluates operational conditions and generates appropriate control signals, acting as a mediator that simplifies the overall control architecture. Rather than having complex control logic directly embedded in each computation portion, the predication logic centralizes the decision-making function, reducing device complexity while achieving power consumption reduction through selective activation.

Inventive Principle:
Principle #24Intermediary (Mediator)

Data Source

PatentUS8191019B2Non-predicated to predicated conversion of asynchronous representations
Publication Date: 2012.05.29 ACHRONIX SEMICONDUCTOR CORP
  • US8191019B2 patent drawing
  • US8191019B2 patent drawing
  • US8191019B2 patent drawing

AI summary

Methods, circuits and systems for converting of a non-predicated asynchronous netlist to a predicated asynchronous netlist are described. These may operate to identify one or more portions of an asynchronous netlist corresponding to a partially utilized portion of an asynchronous circuit. The asynchronous netlist may be modified to control the partially utilized portion. Additional methods, circuits, and systems are disclosed.