Rasterization Band Segmentation for Multi-Core Memory Optimization
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The increasing number of cores in multi-core processors for image forming apparatus leads to a significant increase in memory usage required for rasterization processing, as each band needs a work region proportional to the number of cores, resulting in inefficient memory allocation.
Innovation Solution
The image processing apparatus divides the band into segmented regions based on the number of operation sections, assigning each segmented region to a core for rasterization processing, thereby reducing the memory required for each core's work region and optimizing memory usage.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If rasterization processing is performed by dividing image data into multiple bands and assigning each band to a core, then parallel processing capability is improved, but memory usage increases proportionally with the number of cores
Solution Approach 1:
The patent applies segmentation by dividing each band into multiple sub-bands and assigning different sub-bands to different cores. This allows multiple cores to process portions of the same band simultaneously, improving parallel processing capability while reducing the memory burden on each individual core since each core only needs to hold work regions for its assigned sub-bands rather than entire bands.
2Productivity
If the number of operation sections (cores) is increased to improve processing speed, then productivity increases, but the work region capacity required increases accordingly
Solution Approach 1:
By segmenting bands into sub-bands and distributing them across cores, the patent enables more cores to work in parallel without proportionally increasing total memory usage. Each core processes a smaller portion (sub-band) of the data, reducing individual work region requirements while maintaining high parallel processing throughput.
Solution Approach 2:
The patent implements dynamic work region management where work regions are allocated and released based on the processing progress of each core. As cores complete sub-bands, their work regions become available for reuse by other cores, allowing the system to handle more cores without proportionally increasing peak memory usage.
Data Source
AI summary
The image processing apparatus includes: a storage section including a region to be used when rasterization processing is performed with respect to intermediate language form data created by a band based on page description language form data; a plurality of operation sections; and a control section to perform the rasterization processing with respect to the intermediate language form data by using the operation sections. The control section generates a plurality of segmented regions by dividing the band constituting each pieces of the intermediate language form data into a plurality of pieces based on the number of the operation sections, and assigns a segmented region for executing the rasterization processing among the segmented regions to each of the operation sections.