Multi-Fanout Net Routing via Concurrent Source Wave Expansion
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional routing techniques for programmable ICs, such as FPGAs, perform sub-optimally due to serial processing of multi-fanout nets, leading to increased usage of routing resources and potential timing issues.
Innovation Solution
A method involving the creation and concurrent expansion of source waves, where each wave is assigned a target and expanded based on costs, allowing for dynamic path determination and retention of previous wave expansions to optimize routing without starting anew when a target is coupled.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If conventional serial routing techniques are used to route multi-fanout nets, then routing can be completed with simple processing, but router runtime increases and routing quality deteriorates
Solution Approach 1:
The patent segments the routing problem by creating multiple independent source waves, each responsible for routing to a specific target. Instead of processing the entire multi-fanout net serially, the router divides it into parallel wave expansions, where each wave independently explores paths to its assigned target, thereby reducing overall runtime while maintaining routing quality.
Solution Approach 2:
The patent performs preliminary actions by pre-computing and storing expansion data from previous wave operations. When a target is coupled, the router retains and reuses expansion data from prior waves rather than discarding it, allowing subsequent waves to benefit from previously explored paths and reducing redundant computations, thus improving both speed and quality.
2Reliability
If serial processing of multi-fanout nets is performed, then routing resource usage can be managed simply, but timing cost computation becomes inaccurate and resource usage increases
Solution Approach 1:
The router performs preliminary expansion operations and stores the resulting data for later reuse. By retaining expansion data from previous waves and applying it to subsequent routing operations, the system achieves more accurate timing cost computations without requiring additional routing resources, as the same physical resources are utilized more efficiently across multiple waves.
3Device complexity
If conventional routing techniques are used, then implementation can be simplified, but routing resources are overused and timing critical issues arise
Solution Approach 1:
The patent segments the routing implementation into multiple independent waves, each handling a portion of the multi-fanout net. This segmentation allows the complex task of routing high fanout nets to be divided into manageable, parallel operations that can be executed efficiently, reducing overall resource usage while maintaining implementation feasibility through systematic wave management.
Solution Approach 2:
The router implements a discard-and-recover mechanism where expansion data is retained during wave processing and selectively reused. Instead of discarding expansion data after each wave completion, the system recovers and reuses valuable expansion information from previous waves, thereby reducing the need for additional routing resources while managing implementation complexity through structured data retention.
Data Source
AI summary
Routing a multi-fanout net includes selecting a driver component of the multi-fanout net of a circuit design, wherein the circuit design is specified programmatically, and determining a plurality of targets of the driver component. A source wave is created at each of a plurality of nodes of the driver component. One target is assigned to each source wave. Each source wave is expanded.