Circuit Power Estimation via Partition Probability Distributions

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Conventional circuit simulators face challenges in accurately estimating power consumption, particularly due to difficulties in creating suitable test benches and excessive runtime for cycle accurate simulations, and often rely on simplistic methods that do not account for variance in power consumption over time.

Innovation Solution

The method involves partitioning a circuit design into multiple partitions, each associated with a probability distribution that specifies parameters correlated with power consumption, allowing for the calculation of an output probability distribution indicating power consumption variance over time, using a processor to execute these operations.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If cycle accurate simulation is performed using back annotated netlist, then measurement precision of power consumption is improved, but loss of time increases excessively

Engineering Contradiction:

Improvepower consumption estimation accuracyVSAvoidsimulation runtime

Core Design Contradiction:

Measurement precisionVSLoss of time

Solution Approach 1:

The circuit design is divided into multiple partitions, each associated with a probability distribution. This segmentation allows the simulation to process power consumption estimation in smaller, manageable units rather than requiring a complete cycle accurate simulation of the entire circuit, thereby reducing runtime while maintaining acceptable precision through statistical aggregation of partition results.

Inventive Principle:

Principle #1Segmentation

Solution Approach 2:

Instead of performing actual cycle accurate simulation, the patent uses probability distributions to model and copy the statistical behavior of power consumption. This approach captures the essential characteristics of power variation without requiring the computationally expensive actual simulation, achieving a balance between accuracy and runtime.

Inventive Principle:

Principle #26Copying

2Productivity

If conventional single value estimation method is used, then productivity is improved, but measurement precision of power consumption variance deteriorates

Engineering Contradiction:

Improveestimation speedVSAvoidpower consumption variance characterization

Core Design Contradiction:

ProductivityVSMeasurement precision

Solution Approach 1:

The patent transitions from static single-value estimation to dynamic probability distribution-based estimation. Each partition is associated with a probability distribution that captures the dynamic nature of power consumption, allowing the system to represent power variance and temporal characteristics while maintaining computational efficiency through probabilistic modeling rather than exhaustive simulation.

Inventive Principle:

Principle #15Dynamics

Data Source

PatentUS9268898B1Estimating power consumption of a circuit design

Publication Date: 2016.02.23 XILINX INC

AI summary

Estimating power consumption of a circuit design includes associating, using a processor, each partition of a plurality of partitions of a circuit design with a probability distribution (315). For each partition, the associated probability distribution specifies a distribution for a probability distribution parameter correlated with power consumption for the partition. Using the processor, an output probability distribution specifying power consumption of the circuit design can be calculated according to the probability distribution of each partition of the circuit design (320).