NoC Generation System for Cost-Performance Tradeoff Analysis
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current Network-on-Chip (NoC) designs face challenges in efficiently managing performance and cost, as they often require repeated redesigns to meet varying customer requirements, leading to time and resource wastage, especially when the exact traffic profile is unknown at design time.
Innovation Solution
A method is introduced to process NoC specifications for multiple performance requirements, generating multiple NoC designs that meet minimum performance criteria, and presenting the actual performance and costs of each design, allowing customers to select the most suitable option based on a performance versus cost gradient, utilizing machine learning algorithms to optimize traffic flow mapping and reconfigure NoC hardware.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If multiple NoC designs are generated to meet varying customer requirements, then customer satisfaction and performance matching improve, but design complexity and resource consumption increase
Solution Approach 1:
The system dynamically generates multiple NoC designs based on varying customer requirements and performance criteria. The design generation process adapts to different traffic profiles, performance requirements, and cost constraints, producing optimized NoC configurations on-demand rather than using static pre-defined designs.
Solution Approach 2:
The system varies key parameters such as traffic profile mappings, router configurations, and link allocations to generate diverse NoC designs. By changing these parameters systematically, the system creates multiple viable designs that meet different customer requirements while managing design complexity through parameterized generation.
2Reliability
If NoC redesigns are performed repeatedly to meet changing requirements, then performance requirements are met, but time and resource wastage increase
Solution Approach 1:
The system performs preliminary generation of multiple NoC designs that meet minimum performance requirements before final customer selection. By pre-generating a portfolio of viable designs with different performance-cost tradeoffs, the system avoids repeated redesign cycles and allows customers to select the most suitable option upfront.
Solution Approach 2:
The system creates multiple copies/variations of NoC designs based on a base configuration, modifying parameters to generate diverse options. These copied and adapted designs serve as ready-to-select alternatives, eliminating the need for time-consuming redesigns when customers have different preferences or requirements.
3Loss of information
If detailed performance analysis is provided for each NoC design, then customer decision-making improves, but processing time and computational resources increase
Solution Approach 1:
The system extracts and presents only the most relevant performance metrics and cost information for each generated NoC design, such as bandwidth, latency, power consumption, and area. By selecting and presenting key performance indicators rather than exhaustive analysis, the system enables informed customer decisions without excessive processing time.
Data Source
AI summary
Example implementations as described herein are directed to systems and methods for processing a NoC specification for a plurality of performance requirements of a NoC, and generating a plurality of NoCs, each of the plurality of NoCs meeting a first subset of the plurality of performance requirements. For each of the plurality of NoCs, the example implementations involve presenting a difference between an actual performance of the each of the plurality of NoCs and each performance requirement of a second subset of the plurality of performance requirements and one or more costs for each of the plurality of NoCs.