Functional Primitive Detection in Hardware Netlists
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for automatically detecting functional primitives in hardware systems are inefficient due to coding style variability and exponential complexity, requiring multiple search patterns and failing to detect variations not included in definitions.
Innovation Solution
A method involving mapping netlist representations to a monopartite target graph with hierarchical search patterns, using a pattern graph specification language (PGSL) to abstract variability and efficiently detect functional primitives by disregarding irrelevant details and using constraint satisfaction problems for candidate selection.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Difficulty of detecting and measuring
If direct searching of HDL descriptions is performed to detect functional primitives, then detection capability is provided, but the complexity increases exponentially due to arbitrary coding styles and multiple combination options
Solution Approach 1:
The patent introduces a netlist representation as an intermediary form between HDL descriptions and functional primitive detection. The netlist provides a standardized, language-independent representation that eliminates the exponential complexity arising from arbitrary HDL coding styles while preserving the functional structure needed for detection.
Solution Approach 2:
The patent transforms the detection problem from operating on HDL source code with arbitrary styling parameters to operating on netlists with standardized structural parameters. This parameter transformation converts an intractable search problem into a manageable pattern matching problem.
2Adaptability or versatility
If multiple search patterns are defined to cover coding style variations, then detection coverage is improved, but the number of patterns and search complexity increase
Solution Approach 1:
The patent creates a universal netlist representation that can represent functional primitives regardless of the original HDL coding style. A single search pattern operating on this universal representation can detect multiple variations of functional primitives, eliminating the need for numerous specialized patterns.
Solution Approach 2:
The patent extracts the essential structural characteristics of functional primitives from diverse HDL implementations and encodes them in a single search pattern. This extraction process removes the variability introduced by different coding styles while retaining the functional identity.
3Stability of the object's composition
If netlist representations are used to eliminate coding style discrepancies, then structural equivalence is achieved, but unnecessary details still cause diversity that impedes automatic detection
Solution Approach 1:
The patent extracts only the essential structural elements needed for functional primitive detection from the netlist, removing unnecessary implementation details that create diversity. This selective extraction maintains structural equivalence while reducing irrelevant variation.
Data Source
AI summary
A method for automatic detection of a functional primitive in a model of a hardware system, the model being a netlist having cells and net links therebetween, comprising the steps:a) mapping the cells to target nodes, each of which having a target node type, and the net links to edges of a target graph, and mapping the functional primitive to a search pattern having search nodes and connections therebetween;b) selecting candidates from those target nodes the target node types of which match a search node type, andselecting a candidate structure from those selected candidates the target nodes and edges of which match the search nodes and connections of the search pattern;c) reverse-mapping the target nodes and edges of the selected candidate structure to the cells and net links of the netlist; andd) outputting said cells and net links as detected functional primitive.