Weighted-Cost Interconnect Synthesis for FPGA Design
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current design tools for systems on target devices, such as FPGAs and ASICs, often rely on pre-defined strategies or hard constraints, limiting the exploration of available design options and failing to fully optimize interconnect architectures, especially when incorporating multiple intellectual property (IP) cores with unique interfacing requirements.
Innovation Solution
A method and apparatus for designing a system using weighted-cost interconnect synthesis, where users can select and weight objectives like throughput, latency, resource count, and power consumption to generate an interconnect architecture that optimizes these characteristics, allowing for the generation of candidate solutions that prioritize user-defined goals and constraints.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If pre-defined strategies or hard constraints are applied in design tools, then the design process is simplified and easier to operate, but the exploration of available design options is limited and optimization is reduced
Solution Approach 1:
The patent implements dynamic weighting factors that allow users to adjust the importance of different cost functions (such as area, power, timing) during the design process. This dynamic adjustment enables the design tool to adapt to different design priorities and explore various design spaces, resolving the contradiction between ease of operation and design option exploration by providing both automated guidance and flexible adaptability
2Adaptability or versatility
If multiple IP cores with unique interfacing requirements are incorporated, then system functionality and versatility are improved, but the complexity of interconnect architecture design increases
Solution Approach 1:
The patent changes the parameters of the design process by introducing weighted cost functions that quantify different aspects of interconnect architecture (area, power consumption, timing). By parameterizing the design objectives and using mathematical optimization to balance multiple competing parameters, the tool can handle the complexity of multiple IP cores systematically rather than through manual trial-and-error, thus managing interconnect complexity while maintaining system versatility
3Productivity
If design tools focus on optimizing specific characteristics like throughput or power consumption, then performance in that specific area is improved, but other important characteristics may be compromised
Solution Approach 1:
The patent transforms the design problem into a multi-parameter optimization problem where throughput, power consumption, area, and timing are all represented as cost functions with adjustable weights. This allows the designer to optimize for one parameter (e.g., throughput) while controlling or limiting other parameters (e.g., power consumption) through the weighting mechanism, resolving the contradiction by treating all parameters as simultaneously important but differently weighted
Data Source
AI summary
A method for generating a design for a system implemented on a target device includes presenting a user with an interface that allows the user to weight objectives for an interconnect architecture of the design. The interconnect architecture is generated in response to weighted objectives provided by the user.