Tile-Based Graphics Processing System Transaction Elimination
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Graphics processing systems face significant power consumption and memory bandwidth issues due to frequent writing of data to frame buffers, especially in high-resolution displays, with existing techniques like on-chip frame buffers, caching, and compression having drawbacks such as resource requirements, complexity, and image quality loss.
Innovation Solution
A method where the graphics processing system generates data in blocks (tiles) and compares each block with existing data in the output buffer before writing, determining whether to write based on the comparison, thereby reducing unnecessary transactions and conserving power and bandwidth.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If frame buffer operations are performed frequently to maintain high-resolution display output, then image quality is maintained, but power consumption and memory bandwidth usage increase significantly
Solution Approach 1:
The patent performs a comparison between the current frame buffer data and previous frame data before actually writing to the frame buffer. This preliminary action (comparison) identifies which blocks have changed, allowing the system to skip unnecessary write operations and reduce power consumption while maintaining image quality by only updating changed regions.
Solution Approach 2:
The patent divides the frame buffer into multiple blocks and processes them independently. By segmenting the frame buffer into smaller manageable blocks, the system can compare and update only the necessary blocks rather than the entire frame buffer, reducing overall power consumption and memory bandwidth usage.
2Manufacturing precision
If frame buffer operations are performed frequently to maintain high-resolution display output, then image quality is maintained, but memory bandwidth usage increases significantly
Solution Approach 1:
The comparison operation serves as a preliminary action that identifies which frame buffer blocks require updates. By performing this comparison before writing, the system avoids transferring unnecessary data over the memory bus, thereby reducing memory bandwidth consumption while ensuring that only necessary updates are performed to maintain image quality.
Solution Approach 2:
By dividing the frame buffer into blocks and processing them individually, the patent enables selective updating of only those blocks that have changed. This segmentation approach reduces the total volume of data transferred to and from memory, decreasing memory bandwidth usage while maintaining complete image quality through precise block-level updates.
3Use of energy by moving object
If on-chip frame buffer is used to reduce power consumption, then power usage decreases, but on-chip resource requirements increase significantly
Solution Approach 1:
The patent segments the frame buffer into blocks and implements comparison logic that operates on individual blocks. This segmentation allows the system to use minimal on-chip resources for storing and comparing block data, avoiding the need for a large on-chip frame buffer while still achieving power consumption reduction through selective updating of only changed blocks.
Solution Approach 2:
The patent extracts only the essential comparison function and implements it with minimal on-chip resources. By taking out only the necessary comparison logic and block-level processing capabilities, the system achieves power consumption reduction without requiring significant on-chip resource allocation for a full on-chip frame buffer.
4Use of energy by moving object
If frame buffer compression is used to reduce power consumption and bandwidth usage, then power and bandwidth consumption decrease, but device complexity increases
Solution Approach 1:
The patent performs a simple comparison operation as a preliminary action before frame buffer writing. This comparison is much less complex than compression logic, yet achieves similar power and bandwidth savings by identifying and skipping unchanged blocks, thereby reducing device complexity while maintaining power efficiency.
Solution Approach 2:
Instead of using complex compression logic to reduce data size, the patent uses a simpler approach of copying only the changed blocks to the frame buffer. This block-level copying mechanism achieves power and bandwidth reduction without requiring the complexity of compression algorithms, hardware or software.
5Use of energy by moving object
If dynamic colour depth control is used to reduce power consumption, then power usage decreases, but image quality is reduced due to lossy compression
Solution Approach 1:
The patent performs a preliminary comparison between current and previous frame buffer blocks to identify unchanged regions. This comparison-based approach reduces power consumption by skipping writes to unchanged blocks without altering colour depth or applying lossy compression, thereby maintaining full image quality while achieving power savings through selective updating.
Data Source
AI summary
A transaction elimination hardware unit controls the writing to a frame buffer in a memory of tiles generated by a tile-based graphics processor. The transaction elimination hardware unit has a signature generator that generates a signature representative of the content of the tile for each tile. A signature comparator then compares the signature of a new tile received from the graphics processor with the signatures of one or more tiles already stored in the frame buffer to see if the signatures match. If the signatures do not match, then the signature comparator controls a write controller to write the new tile to the frame buffer. On the other hand, if the signatures match, then no data is written to the frame buffer and the existing tile is allowed to remain in the frame buffer. In this way, a tile is only written to the frame buffer if it is found by the signature comparison to differ from the tile or tiles that are already stored in the frame buffer that it is compared with.