Circuit Power Measurement Using Toggle Windows and Weighting
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current power analysis flows at the RTL level are slow and unscalable due to computationally expensive simulations and reliance on large-scale benchmarks, making it challenging to perform per-cycle power simulation in large digital electronic circuits.
Innovation Solution
A data-driven power modeling approach that organizes toggle data into time windows and uses a weighting network to combine averaged toggle data from selected power proxy signals, allowing for simultaneous signal selection and power model calibration, thereby reducing the complexity of large-scale circuits and enhancing simulator-based and emulator-based solutions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If computer simulations are used to determine power usage in circuits with hundreds of thousands of registers, then power measurement accuracy is improved, but simulation time increases significantly
Solution Approach 1:
The patent segments the power measurement problem by dividing the circuit into multiple clock cycles and processing them in batches. The simulator executes workloads and collects toggle data for multiple clock cycles, then processes this data in organized time windows rather than simulating each cycle individually. This segmentation enables parallel processing and significantly reduces the overall simulation time while maintaining measurement accuracy.
Solution Approach 2:
The patent performs preliminary actions by pre-organizing toggle data into time windows and pre-processing signal data before final power calculation. The system collects toggle data for multiple clock cycles in advance, organizes it into structured time windows, and prepares averaged toggle data before the actual power measurement computation. This preliminary organization reduces the computational burden during the final measurement phase.
2Measurement precision
If all signal toggle data is processed for power calculation, then measurement completeness is improved, but computational complexity increases
Solution Approach 1:
The patent extracts only the essential information needed for power calculation by focusing on toggle data organization and averaging. Instead of processing all raw signal data individually, the system extracts key features by averaging toggle data within time windows and selecting representative signals. This extraction approach maintains measurement completeness while significantly reducing computational complexity.
Solution Approach 2:
The patent applies partial action by processing a representative subset of data through averaging and time-windowing operations. Rather than performing exhaustive calculations on every single signal toggle, the system processes averaged data from time windows, which provides sufficient measurement completeness for power analysis while avoiding the excessive computational burden of processing every individual signal event.
3Measurement precision
If per-cycle power simulation is performed in large digital electronic circuits, then power analysis accuracy is improved, but scalability deteriorates
Solution Approach 1:
The patent segments the large circuit analysis into manageable time windows and batches of clock cycles. By dividing the simulation into organized time windows and processing signals in structured batches, the system maintains per-cycle power analysis accuracy for large circuits while improving scalability through efficient resource utilization and parallel processing capabilities.
Solution Approach 2:
The patent performs preliminary data collection and organization for multiple clock cycles before final power calculation. By pre-executing simulations for multiple cycles, pre-organizing toggle data into time windows, and pre-computing averaged values, the system enables scalable per-cycle power analysis in large circuits without the computational burden of real-time processing.
Data Source
AI summary
A power meter for measuring power usage in a circuit includes preprocessor and a weighting network. The pre-processor is configured to receive toggle data for a number of power proxy signals in the circuit for a plurality of clock cycles of the circuit in a first time window. The power proxy signals and weighting values are determined automatically from simulated or emulated toggle data. For each power proxy signal, the pre-processor averages the toggle data over one or more clock cycles in one or more second time windows, within the first time window, to provide averaged toggle data, and outputs the averaged toggle data for each second time window. The weighting network is configured to combine the averaged toggle data from the power proxy signals, based on a set of weight values, to provide a measure of the power usage.