IP Block Power Profile via Node Sampling and Net Weighting
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for generating a power profile of electronic circuits are expensive and time-consuming, requiring extensive simulation and data gathering, especially when dealing with incomplete or missing toggle simulation data, which affects the accuracy of power mode characterization.
Innovation Solution
The method employs node sampling techniques to estimate the power profile by selecting populated nets with toggle simulation data, generating a sample energy, and using a ratio of net weights to model a block power profile, thereby reducing the need for extensive simulations and accounting for energy expenditure on nets lacking toggle data.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If extensive simulation and data gathering methods are used to generate power profiles, then measurement precision is improved, but loss of time and productivity deteriorate
Solution Approach 1:
The patent applies partial action by performing simulation only on a selected subset of nets (populated nets with toggle data) rather than all nets in the design. The method identifies nets that have toggle simulation data and uses only those for power profile generation, thereby reducing simulation time while maintaining acceptable accuracy through the weighting mechanism that accounts for unsimulated nets
2Measurement precision
If extensive simulation and data gathering methods are used to generate power profiles, then measurement precision is improved, but productivity deteriorates
Solution Approach 1:
The method performs power analysis on only the populated nets that have toggle simulation data available, rather than requiring complete simulation of all nets. This partial approach significantly reduces the computational burden and accelerates design iterations while the weighting mechanism compensates for the missing data from unsimulated nets
3Productivity
If node sampling techniques are used to estimate power profiles, then productivity is improved, but measurement precision deteriorates
Solution Approach 1:
The patent transforms the sampling approach by introducing net weights that are proportional to effective load capacitance. This parameter change allows the method to account for the relative importance and energy contribution of each net, thereby improving the accuracy of the estimated power profile generated from the sampled populated nets
Solution Approach 2:
The weighting mechanism acts as an intermediary that bridges the gap between the sampled populated nets and the complete set of nets. By applying weights proportional to load capacitance, the method indirectly accounts for the energy expenditure of unsimulated nets, improving accuracy without requiring direct simulation of all nets
4Device complexity
If node sampling techniques are used to estimate power profiles, then processing requirements are reduced, but measurement precision deteriorates
Solution Approach 1:
The method changes the parameter of net selection from uniform sampling to weighted sampling based on effective load capacitance. This parameter change ensures that nets with higher capacitance (which contribute more to power consumption) are given appropriate weight in the estimation, maintaining accuracy while reducing processing requirements by sampling only populated nets
Data Source
AI summary
A power profile for an electronics design is implemented by accessing a unit descriptive of an electronic design comprising a first intellectual property (IP) block expressed in a simulation language and comprising a netlist. A total number (NT) of net weights are identified in the netlist, wherein each respective net weight is proportional to an effective load capacitance of an associated net. A total number (NP) of populated nets having associated toggle simulation data are identified in the netlist. A ratio (KS) equal to a sum of all NT net weights divided by a sum of all NP populated net weights is generated. A sample energy (ES) is generated based on the associated toggle simulation data of and net weights for each of the NP populated nets. And a block power profile is modelled based on an estimated block energy (EN) equal to KS multiplied by ES.