2D Transform Zero-Value Bypass for Lower-Power Processing
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional methods for performing 2D transforms on data arrays with significant zero values are inefficient, leading to high power consumption and performance issues, particularly in mobile devices that require low power consumption and stringent processing cycles.
Innovation Solution
The system performs 2D transforms by determining subsets of rows and columns with zero values and bypassing or reducing power consumption for these operations, using circuitry to identify and manage zero values within the transform engine, allowing for reduced power usage and improved performance by avoiding unnecessary computations.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of energy
If conventional 2D transform methods are used on data arrays with significant zero values, then complete transform operations are performed on all data values, but power consumption increases and computational efficiency decreases
Solution Approach 1:
The transform engine divides the 2D data array into multiple rows and columns, and further segments them into subsets. Zero-value detection is performed on each subset independently, allowing the system to identify and bypass transformations for zero-value subsets while processing only non-zero subsets, thereby reducing overall power consumption and improving computational efficiency
Solution Approach 2:
Instead of performing complete transform operations on all data values, the system performs partial actions by detecting zero values and bypassing transformation operations for those specific elements. The transform engine selectively applies transformation only to non-zero data values, avoiding unnecessary computations and reducing power consumption
2Reliability
If transform operations are performed on all data values including zero values, then complete transformation is achieved, but redundant operations increase power consumption
Solution Approach 1:
The system extracts and identifies zero values from the data array using dedicated zero-value detection circuitry. Once zero values are extracted and identified, the transform engine bypasses transformation operations for these elements, performing transformations only on non-zero values. This ensures transform accuracy is maintained for all non-zero elements while eliminating redundant operations on zero values that would consume power
3Productivity
If full transform operations are performed on each row and column, then complete 2D transform is achieved, but processing cycles increase and performance decreases
Solution Approach 1:
The system performs preliminary zero-value detection on rows and columns before executing transform operations. By detecting zero values in advance and marking them for bypass, the transform engine can skip unnecessary transformation steps during the actual processing phase, significantly reducing processing cycles and improving processing speed without compromising the completeness of the transform for non-zero elements
Data Source
AI summary
A method and system for performing a 2D transform is disclosed. The 2D transform may include a row transform and/or a column transform. When performing the row or column transform, it may be determined whether each of different subsets of the data values including a partition of a row or column includes at least one zero value, whether each of different subsets of a first subset of the partition includes at least one zero value, and whether each of different subsets of at least one other subset of the partition includes at least one zero value. When performing the row or column transform, at least one transformation operation on at least one zero value may be bypassed or performed in a reduced-power manner, where such transformation operation would otherwise be performed in a manner consuming full power if the zero value were a non-zero value.