Mask Data Preparation Branching for Parallel Processing
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In integrated circuit design and manufacturing, the shrinking feature sizes increase the difficulty of faithfully reproducing the intended layout design onto the substrate, leading to pattern-dependent defects and increased computational costs due to serial design processes, which hinder the parallelization of data preparation operations.
Innovation Solution
The method involves branching the data-flow in mask data preparation processes by connecting the output stream of a first mask data processing operation to independent mask data preparation operations, allowing them to operate in parallel on discrete portions of the data, thereby facilitating section-mode processing and reducing computational overhead.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If serial mask data preparation processes are used, then computational accuracy is maintained, but processing time increases and productivity decreases
Solution Approach 1:
The patent divides the mask data preparation process into multiple independent sections or modules that can be processed separately. Each section handles specific portions of the mask data independently, allowing parallel execution while maintaining overall process accuracy. This segmentation enables the system to process different regions of the mask simultaneously rather than sequentially.
Solution Approach 2:
The patent performs preliminary analysis and classification of mask data to identify independent processing sections before the actual mask data preparation begins. By pre-processing the data to determine which sections can be handled independently, the system prepares the groundwork for parallel processing, reducing the time required for the main preparation operations while ensuring accuracy requirements are met.
2Productivity
If feature sizes are shrunk to increase circuit density, then more circuits can be integrated, but pattern-dependent defects increase and manufacturing precision becomes more difficult to maintain
Solution Approach 1:
The patent applies different processing qualities and parameters to different local regions of the mask data based on their specific characteristics. Instead of using a uniform processing approach, the system identifies regions with different feature sizes, densities, and complexity, and applies optimized processing parameters to each region. This local quality approach ensures that each region receives the appropriate level of processing attention needed to maintain pattern fidelity despite varying feature dimensions.
Solution Approach 2:
The patent incorporates feedback mechanisms that monitor the processing results and adjust parameters in real-time to compensate for pattern-dependent defects. By continuously analyzing the output of each processing section and comparing it against desired specifications, the system can identify and correct deviations, ensuring that even as feature sizes shrink, the final mask data maintains the required precision for faithful pattern reproduction.
Data Source
AI summary
Branching of the data-flow in a mask data preparation processes is described herein. In various implementations, the output stream from a first mask data processing operation is branched. Subsequently, the branched output stream may be connected to the input stream of a first independent mask data preparation operation and a second independent mask data preparation operation. This provides that the first and the second independent mask data preparation operations may operate in parallel. Furthermore, this provides that the first and the second independent mask data preparation operations may operate upon discrete “portions” of the data processed by the first mask data preparation operation.