Single-Fix Rectification Function for Netlist Functional Equivalence
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional integrated circuit design systems are inefficient in determining changes to an original netlist to achieve functional equivalence with a target netlist, especially when dealing with complex logical differences, and often require costly re-generation of gate masks.
Innovation Solution
A process is developed to identify and modify single-fix signals in the original netlist to derive rectification functions, allowing for functional equivalence with the target netlist without re-executing logic synthesis, placement, or routing design processes, and utilizing existing gate masks to minimize fabrication costs.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If conventional design tools are used to determine placement and routing based on netlist changes, then functional equivalence can be achieved, but gate masks must be regenerated at significant cost
Solution Approach 1:
The invention extracts and modifies only the specific signal(s) that need to be changed to achieve functional equivalence, rather than regenerating the entire netlist and associated gate masks. This selective extraction approach allows changing the logical function while reusing existing fabrication masks, significantly reducing manufacturing costs.
Solution Approach 2:
The invention changes the logical parameters of specific signals (e.g., changing signal S1 from (c^d)' to (c^d)' + e') to achieve functional equivalence. By modifying only the necessary signal parameters rather than the entire design, the system achieves the desired functional change without requiring costly mask regeneration.
2Reliability
If logic synthesis is re-executed to implement changes, then functional equivalence can be achieved, but engineering effort and time are significantly increased
Solution Approach 1:
The invention performs preliminary identification of single-fix signals and derives rectification functions before implementing changes. By preparing the rectification logic in advance based on the logical difference between original and target netlists, the system achieves functional equivalence without requiring time-consuming re-execution of logic synthesis, placement, and routing processes.
3Adaptability or versatility
If conventional systems locate single-fix signals and derive rectification functions, then functional changes can be implemented, but the systems are not scalable and are limited to simple correcting models
Solution Approach 1:
The invention provides a universal method for determining single-fix signals and deriving rectification functions that works across different types of logical differences and netlist configurations. The system is not limited to simple correcting models but can handle complex functional changes by identifying signals whose modification achieves functional equivalence between original and target netlists.
Data Source
AI summary
Some aspects provide determination of a function to rectify functional differences between netlist G1 and netlist G2 having inputs V. The determination may include determination of a signal s of netlist G1 that can be re-synthesized so as to correct the functional differences between netlist G1 and netlist G2, assignment of respective static values to a first plurality of inputs V, assignment of respective initial values to a second plurality of inputs V, determination of a first function based on the assigned static values, the assigned initial values, a first error function reflecting the difference between outputs of netlist G1 and netlist G2 for each vector of inputs V in a case that s equals 0, and a second error function reflecting the difference between the outputs of netlist G1 and netlist G2 for each vector of inputs V in a case that s equals 1. Also included may be determination of whether the first function rectifies the functional differences between netlist G1 and netlist G2, assignment, if it is determined that the first function does not rectify the functional differences, of respective next values to the second plurality of inputs, and determination of a second function based on the first function, the assigned static values, the assigned next values, the first error function, and the second error function.