Memory-Mapped Disk for IC Mask Layout Import
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The existing methods for importing mask layout data into a target system are time-consuming due to the large size of hierarchical layouts, leading to increased productivity issues, as they require extensive CPU and I/O operations, and disk sharing among multiple processes and users.
Innovation Solution
The use of a memory-mapped disk with a 'shadow' hierarchy for intermediate translation during data import, allowing for faster I/O storage access and reduced need for disk access, by translating mask layout data into cell views and transmitting them from the memory-mapped disk to the target system's magnetic disk.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of time
If mask layout data is imported using traditional disk-based methods, then data can be stored and accessed, but the import process becomes extremely time-consuming due to large data sizes and extensive I/O operations
Solution Approach 1:
The patent introduces a memory-mapped disk as an intermediary storage layer between the host system and the target system. This intermediary uses RAM to store mask layout data during the import process, eliminating the need for repeated disk I/O operations while maintaining data accessibility. The memory-mapped disk acts as a buffer that decouples the CPU-intensive translation process from the I/O-bound disk operations, thereby resolving the time loss and productivity issues.
Solution Approach 2:
The patent transitions from traditional disk-based storage (slow, sequential access) to memory-mapped storage (fast, random access) by utilizing the RAM address space. This dimensional change in storage medium fundamentally alters the access characteristics, enabling parallel CPU and disk operations without the bottlenecks of traditional sequential disk I/O, thus dramatically reducing import time and improving productivity.
2Ease of manufacture
If CPU-intensive translation operations are performed, then layout data can be converted to target format, but disk access operations become interleaved and time-consuming
Solution Approach 1:
The memory-mapped disk serves as an intermediary that decouples CPU translation operations from disk access operations. By mapping the target disk location into the host system's address space, the system can perform CPU-intensive translation while simultaneously accessing data from the memory-mapped disk without interleaving delays. This eliminates the time loss caused by context switching between CPU and disk operations.
Solution Approach 2:
The patent enables continuous translation operations by providing the CPU with uninterrupted access to layout data through the memory-mapped disk. The RAM-based storage ensures that data remains readily accessible during translation without requiring disk seek operations, maintaining continuous useful action throughout the translation process and significantly reducing overall translation time.
3Adaptability or versatility
If target disk is shared among multiple processes and users, then resource utilization is improved, but disk access speed decreases due to head movements and latency
Solution Approach 1:
The memory-mapped disk acts as an intermediary buffer that absorbs the impact of shared disk access. Multiple processes and users can access the same memory-mapped region simultaneously without causing disk head movements, as all access occurs in RAM. This intermediary layer maintains the adaptability of disk sharing while eliminating the speed degradation caused by concurrent disk access conflicts.
Solution Approach 2:
The patent creates a memory-based copy of the target disk location that can be accessed by multiple processes and users simultaneously. This copy resides in RAM and provides fast, concurrent access without the limitations of physical disk sharing. The copying principle allows unlimited parallel access to the same data without the mechanical constraints of disk head movements, maintaining versatility while achieving high speed.
Data Source
AI summary
A host system for transferring data to a target system is provided. The host system may include a layout database for storing mask layout data representing an integrated circuit (IC) in terms of planar geometric shapes. The hosts system may further include a processor configured to import the mask layout data from the layout database to a memory-mapped disk in the host system. The processor is further configured to translate the mask layout data into one or more cell views according to a table hierarchy in the memory-mapped disk. The processor is further configured to transmit the one or more cell views from the memory-mapped disk to a magnetic disk of the target system.